{
  "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": 0.012647395953536034,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.0686,
      "num_tokens": 264685.0,
      "reward": 0.037574999034404755,
      "reward_std": 0.07449960708618164,
      "rewards/accuracy_reward_step": 0.015625,
      "rewards/final_brier_reward_step": 0.01655624993145466,
      "rewards/format_reward_step": 0.04296875,
      "step": 1
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.5604067397745032,
      "aux_distill/mean_u": 0.29356464889172007,
      "aux_distill/n_active_tok": 58.36842105263158,
      "calib/answer_extract_rate": 0.13671875,
      "calib/auroc": 0.5338345864661654,
      "calib/avg_num_step_conf": 0.55078125,
      "calib/ece": 0.6261538461538463,
      "calib/final_conf_rate": 0.1015625,
      "calib/format_rate": 0.08984375,
      "calib/frac_conf_gt_0.9": 0.7692307692307693,
      "calib/gap": 0.002406015037593856,
      "calib/mean_conf": 0.8953846153846153,
      "calib/mu_c": 0.897142857142857,
      "calib/mu_w": 0.8947368421052632,
      "calib/nonempty_final_conf_rate": 0.1015625,
      "calib/nonempty_reasoning_rate": 0.14453125,
      "calib/nonempty_step_conf_rate": 0.109375,
      "calib/pce": 0.6261538461538463,
      "calib/std_conf": 0.18653172073466937,
      "calib/step_conf_rate": 0.109375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0546875,
      "completions/max_length": 3001.0,
      "completions/max_terminated_length": 3001.0,
      "completions/mean_length": 646.4609375,
      "completions/mean_terminated_length": 683.8594970703125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0021333333333333334,
      "grad_norm": 0.008666743524372578,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.099,
      "num_tokens": 533467.0,
      "reward": 0.07537207007408142,
      "reward_std": 0.14035090804100037,
      "rewards/accuracy_reward_step": 0.03125,
      "rewards/final_brier_reward_step": 0.02965039201080799,
      "rewards/format_reward_step": 0.08984375,
      "step": 2
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.6924514323472977,
      "aux_distill/mean_u": 0.24409276830107252,
      "aux_distill/n_active_tok": 51.2,
      "calib/answer_extract_rate": 0.0625,
      "calib/auroc": 0.5520833333333334,
      "calib/avg_num_step_conf": 0.26953125,
      "calib/ece": 0.5471428571428572,
      "calib/final_conf_rate": 0.0546875,
      "calib/format_rate": 0.02734375,
      "calib/frac_conf_gt_0.9": 0.8571428571428571,
      "calib/gap": 0.0029166666666665675,
      "calib/mean_conf": 0.9600000000000001,
      "calib/mu_c": 0.9616666666666666,
      "calib/mu_w": 0.95875,
      "calib/nonempty_final_conf_rate": 0.0546875,
      "calib/nonempty_reasoning_rate": 0.0703125,
      "calib/nonempty_step_conf_rate": 0.0546875,
      "calib/pce": 0.5392857142857143,
      "calib/std_conf": 0.03229329872987803,
      "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": 0.08203125,
      "completions/max_length": 3035.0,
      "completions/max_terminated_length": 3035.0,
      "completions/mean_length": 672.12890625,
      "completions/mean_terminated_length": 732.1914672851562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0032,
      "grad_norm": 0.0068152653984725475,
      "learning_rate": 7.5e-07,
      "loss": 0.0357,
      "num_tokens": 810788.0,
      "reward": 0.0320499986410141,
      "reward_std": 0.07596391439437866,
      "rewards/accuracy_reward_step": 0.0234375,
      "rewards/final_brier_reward_step": 0.013318749144673347,
      "rewards/format_reward_step": 0.02734375,
      "step": 3
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.775890588760376,
      "aux_distill/mean_u": 0.24166433444907634,
      "aux_distill/n_active_tok": 37.333333333333336,
      "calib/answer_extract_rate": 0.0703125,
      "calib/avg_num_step_conf": 0.16796875,
      "calib/ece": 0.9345454545454546,
      "calib/final_conf_rate": 0.04296875,
      "calib/format_rate": 0.02734375,
      "calib/frac_conf_gt_0.9": 0.8181818181818182,
      "calib/mean_conf": 0.9345454545454543,
      "calib/mu_c": NaN,
      "calib/mu_w": 0.9345454545454543,
      "calib/nonempty_final_conf_rate": 0.04296875,
      "calib/nonempty_reasoning_rate": 0.08203125,
      "calib/nonempty_step_conf_rate": 0.04296875,
      "calib/pce": 0.9345454545454546,
      "calib/std_conf": 0.0909908723058264,
      "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": 0.078125,
      "completions/max_length": 3033.0,
      "completions/max_terminated_length": 3033.0,
      "completions/mean_length": 621.32421875,
      "completions/mean_terminated_length": 673.9788208007812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 4.0,
      "epoch": 0.004266666666666667,
      "grad_norm": 0.006891297642141581,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0315,
      "num_tokens": 1076015.0,
      "reward": 0.014284765347838402,
      "reward_std": 0.02579544484615326,
      "rewards/accuracy_reward_step": 0.0,
      "rewards/final_brier_reward_step": 0.001225781161338091,
      "rewards/format_reward_step": 0.02734375,
      "step": 4
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.6852438772718111,
      "aux_distill/mean_u": 0.16246693256067343,
      "aux_distill/n_active_tok": 26.666666666666668,
      "calib/answer_extract_rate": 0.0546875,
      "calib/auroc": 0.5357142857142857,
      "calib/avg_num_step_conf": 0.08203125,
      "calib/ece": 0.6766666666666665,
      "calib/final_conf_rate": 0.03515625,
      "calib/format_rate": 0.015625,
      "calib/frac_conf_gt_0.9": 0.5555555555555556,
      "calib/gap": 0.0014285714285714457,
      "calib/mean_conf": 0.8988888888888888,
      "calib/mu_c": 0.9,
      "calib/mu_w": 0.8985714285714286,
      "calib/nonempty_final_conf_rate": 0.03515625,
      "calib/nonempty_reasoning_rate": 0.06640625,
      "calib/nonempty_step_conf_rate": 0.03125,
      "calib/pce": 0.6766666666666665,
      "calib/std_conf": 0.1257373315416304,
      "calib/step_conf_rate": 0.03125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3022.0,
      "completions/max_terminated_length": 3022.0,
      "completions/mean_length": 677.96484375,
      "completions/mean_terminated_length": 754.6043090820312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.005333333333333333,
      "grad_norm": 0.006530698388814926,
      "learning_rate": 1.25e-06,
      "loss": 0.0432,
      "num_tokens": 1356262.0,
      "reward": 0.017167968675494194,
      "reward_std": 0.048558346927165985,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/final_brier_reward_step": 0.010898437350988388,
      "rewards/format_reward_step": 0.015625,
      "step": 5
    },
    {
      "aux_distill/lambda": 0.1,
      "aux_distill/loss": 0.6688954555071317,
      "aux_distill/mean_u": 0.35205855050102514,
      "aux_distill/n_active_tok": 54.76923076923077,
      "calib/answer_extract_rate": 0.08984375,
      "calib/auroc": 0.34615384615384615,
      "calib/avg_num_step_conf": 0.34765625,
      "calib/ece": 0.725,
      "calib/final_conf_rate": 0.0625,
      "calib/format_rate": 0.046875,
      "calib/frac_conf_gt_0.9": 0.6875,
      "calib/gap": -0.021538461538461617,
      "calib/mean_conf": 0.8775000000000001,
      "calib/mu_c": 0.86,
      "calib/mu_w": 0.8815384615384616,
      "calib/nonempty_final_conf_rate": 0.0625,
      "calib/nonempty_reasoning_rate": 0.10546875,
      "calib/nonempty_step_conf_rate": 0.07421875,
      "calib/pce": 0.7075,
      "calib/std_conf": 0.14813422967025547,
      "calib/step_conf_rate": 0.07421875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.08203125,
      "completions/max_length": 3020.0,
      "completions/max_terminated_length": 3020.0,
      "completions/mean_length": 575.47265625,
      "completions/mean_terminated_length": 626.8978271484375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 5.0,
      "epoch": 0.0064,
      "grad_norm": 0.006988861598074436,
      "learning_rate": 1.5e-06,
      "loss": 0.0504,
      "num_tokens": 1609535.0,
      "reward": 0.03744082152843475,
      "reward_std": 0.07511966675519943,
      "rewards/accuracy_reward_step": 0.01171875,
      "rewards/final_brier_reward_step": 0.01628788933157921,
      "rewards/format_reward_step": 0.046875,
      "step": 6
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.6855257414281368,
      "aux_distill/mean_u": 0.32664575830852394,
      "aux_distill/n_active_tok": 36.5625,
      "calib/answer_extract_rate": 0.08203125,
      "calib/auroc": 0.7727272727272727,
      "calib/avg_num_step_conf": 0.30078125,
      "calib/ece": 0.7469230769230768,
      "calib/final_conf_rate": 0.05078125,
      "calib/format_rate": 0.0390625,
      "calib/frac_conf_gt_0.9": 0.7692307692307693,
      "calib/gap": 0.10545454545454536,
      "calib/mean_conf": 0.9007692307692309,
      "calib/mu_c": 0.99,
      "calib/mu_w": 0.8845454545454546,
      "calib/nonempty_final_conf_rate": 0.05078125,
      "calib/nonempty_reasoning_rate": 0.11328125,
      "calib/nonempty_step_conf_rate": 0.0703125,
      "calib/pce": 0.7469230769230768,
      "calib/std_conf": 0.16202856056223489,
      "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": 0.09765625,
      "completions/max_length": 3070.0,
      "completions/max_terminated_length": 3070.0,
      "completions/mean_length": 662.6015625,
      "completions/mean_terminated_length": 734.3117065429688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.007466666666666667,
      "grad_norm": 0.007448374293744564,
      "learning_rate": 1.75e-06,
      "loss": 0.0134,
      "num_tokens": 1886585.0,
      "reward": 0.030934961512684822,
      "reward_std": 0.06625574827194214,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/final_brier_reward_step": 0.01499492209404707,
      "rewards/format_reward_step": 0.0390625,
      "step": 7
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.7116753607988358,
      "aux_distill/mean_u": 0.26608393808950553,
      "aux_distill/n_active_tok": 65.91666666666667,
      "calib/answer_extract_rate": 0.06640625,
      "calib/auroc": 0.16666666666666669,
      "calib/avg_num_step_conf": 0.375,
      "calib/ece": 0.7890909090909093,
      "calib/final_conf_rate": 0.04296875,
      "calib/format_rate": 0.03125,
      "calib/frac_conf_gt_0.9": 1.0,
      "calib/gap": -0.013333333333333308,
      "calib/mean_conf": 0.970909090909091,
      "calib/mu_c": 0.96,
      "calib/mu_w": 0.9733333333333333,
      "calib/nonempty_final_conf_rate": 0.04296875,
      "calib/nonempty_reasoning_rate": 0.08203125,
      "calib/nonempty_step_conf_rate": 0.06640625,
      "calib/pce": 0.7890909090909093,
      "calib/std_conf": 0.011642044068059734,
      "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": 0.08203125,
      "completions/max_length": 2986.0,
      "completions/max_terminated_length": 2986.0,
      "completions/mean_length": 660.27734375,
      "completions/mean_terminated_length": 719.2808227539062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.008533333333333334,
      "grad_norm": 0.005775726865977049,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.0424,
      "num_tokens": 2162128.0,
      "reward": 0.025961915031075478,
      "reward_std": 0.061920247972011566,
      "rewards/accuracy_reward_step": 0.01171875,
      "rewards/final_brier_reward_step": 0.008955078199505806,
      "rewards/format_reward_step": 0.03125,
      "step": 8
    },
    {
      "aux_distill/lambda": 0.09999999999999999,
      "aux_distill/loss": 0.6978093602440574,
      "aux_distill/mean_u": 0.19239018482123108,
      "aux_distill/n_active_tok": 25.454545454545453,
      "calib/answer_extract_rate": 0.0703125,
      "calib/auroc": 0.65,
      "calib/avg_num_step_conf": 0.140625,
      "calib/ece": 0.7108333333333334,
      "calib/final_conf_rate": 0.046875,
      "calib/format_rate": 0.0390625,
      "calib/frac_conf_gt_0.9": 0.6666666666666666,
      "calib/gap": 0.08700000000000008,
      "calib/mean_conf": 0.8775,
      "calib/mu_c": 0.95,
      "calib/mu_w": 0.8629999999999999,
      "calib/nonempty_final_conf_rate": 0.046875,
      "calib/nonempty_reasoning_rate": 0.078125,
      "calib/nonempty_step_conf_rate": 0.05078125,
      "calib/pce": 0.7108333333333334,
      "calib/std_conf": 0.15562374497485915,
      "calib/step_conf_rate": 0.05078125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.1171875,
      "completions/max_length": 2826.0,
      "completions/max_terminated_length": 2826.0,
      "completions/mean_length": 633.48828125,
      "completions/mean_terminated_length": 717.5796508789062,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 4.0,
      "epoch": 0.0096,
      "grad_norm": 0.0072958157397806644,
      "learning_rate": 2.25e-06,
      "loss": 0.0397,
      "num_tokens": 2431837.0,
      "reward": 0.030364451929926872,
      "reward_std": 0.05786587670445442,
      "rewards/accuracy_reward_step": 0.0078125,
      "rewards/final_brier_reward_step": 0.013853906653821468,
      "rewards/format_reward_step": 0.0390625,
      "step": 9
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.6190562213168425,
      "aux_distill/mean_u": 0.2795340197168761,
      "aux_distill/n_active_tok": 38.35294117647059,
      "calib/answer_extract_rate": 0.08203125,
      "calib/auroc": 0.4210526315789474,
      "calib/avg_num_step_conf": 0.31640625,
      "calib/ece": 0.8112499999999999,
      "calib/final_conf_rate": 0.078125,
      "calib/format_rate": 0.05078125,
      "calib/frac_conf_gt_0.9": 0.65,
      "calib/gap": 0.09342105263157874,
      "calib/mean_conf": 0.8612500000000001,
      "calib/mu_c": 0.95,
      "calib/mu_w": 0.8565789473684212,
      "calib/nonempty_final_conf_rate": 0.078125,
      "calib/nonempty_reasoning_rate": 0.11328125,
      "calib/nonempty_step_conf_rate": 0.09375,
      "calib/pce": 0.8112499999999999,
      "calib/std_conf": 0.23361225888210574,
      "calib/step_conf_rate": 0.09375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.078125,
      "completions/max_length": 3041.0,
      "completions/max_terminated_length": 3041.0,
      "completions/mean_length": 744.4375,
      "completions/mean_terminated_length": 807.5254516601562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.010666666666666666,
      "grad_norm": 0.007861468009650707,
      "learning_rate": 2.5e-06,
      "loss": 0.0948,
      "num_tokens": 2729213.0,
      "reward": 0.035783737897872925,
      "reward_std": 0.09035970270633698,
      "rewards/accuracy_reward_step": 0.00390625,
      "rewards/final_brier_reward_step": 0.016879979521036148,
      "rewards/format_reward_step": 0.05078125,
      "step": 10
    },
    {
      "aux_distill/lambda": 0.10000000000000002,
      "aux_distill/loss": 0.6934706956148148,
      "aux_distill/mean_u": 0.29880514015963444,
      "aux_distill/n_active_tok": 48.1,
      "calib/answer_extract_rate": 0.1015625,
      "calib/auroc": 0.5069444444444444,
      "calib/avg_num_step_conf": 0.46875,
      "calib/ece": 0.6405555555555555,
      "calib/final_conf_rate": 0.0703125,
      "calib/format_rate": 0.06640625,
      "calib/frac_conf_gt_0.9": 0.6666666666666666,
      "calib/gap": -0.13750000000000007,
      "calib/mean_conf": 0.785,
      "calib/mu_c": 0.6933333333333334,
      "calib/mu_w": 0.8308333333333334,
      "calib/nonempty_final_conf_rate": 0.0703125,
      "calib/nonempty_reasoning_rate": 0.14453125,
      "calib/nonempty_step_conf_rate": 0.125,
      "calib/pce": 0.5461111111111112,
      "calib/std_conf": 0.3197785692214744,
      "calib/step_conf_rate": 0.125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.08203125,
      "completions/max_length": 2792.0,
      "completions/max_terminated_length": 2792.0,
      "completions/mean_length": 662.73828125,
      "completions/mean_terminated_length": 721.961669921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.011733333333333333,
      "grad_norm": 0.01090265903621912,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.0945,
      "num_tokens": 3003354.0,
      "reward": 0.060478322207927704,
      "reward_std": 0.11573203653097153,
      "rewards/accuracy_reward_step": 0.02734375,
      "rewards/final_brier_reward_step": 0.02720664069056511,
      "rewards/format_reward_step": 0.06640625,
      "step": 11
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.6174011901021004,
      "aux_distill/mean_u": 0.3561512000879931,
      "aux_distill/n_active_tok": 80.75,
      "calib/answer_extract_rate": 0.19921875,
      "calib/auroc": 0.5722222222222222,
      "calib/avg_num_step_conf": 0.94921875,
      "calib/ece": 0.6661538461538461,
      "calib/final_conf_rate": 0.15234375,
      "calib/format_rate": 0.109375,
      "calib/frac_conf_gt_0.9": 0.7435897435897436,
      "calib/gap": 0.07766666666666666,
      "calib/mean_conf": 0.8969230769230768,
      "calib/mu_c": 0.9566666666666666,
      "calib/mu_w": 0.8789999999999999,
      "calib/nonempty_final_conf_rate": 0.15234375,
      "calib/nonempty_reasoning_rate": 0.23828125,
      "calib/nonempty_step_conf_rate": 0.171875,
      "calib/pce": 0.6661538461538461,
      "calib/std_conf": 0.16687403273364296,
      "calib/step_conf_rate": 0.171875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.08203125,
      "completions/max_length": 2993.0,
      "completions/max_terminated_length": 2993.0,
      "completions/mean_length": 604.328125,
      "completions/mean_terminated_length": 658.3319091796875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.0128,
      "grad_norm": 0.009621622040867805,
      "learning_rate": 3e-06,
      "loss": 0.0646,
      "num_tokens": 3262238.0,
      "reward": 0.0920531302690506,
      "reward_std": 0.18517741560935974,
      "rewards/accuracy_reward_step": 0.03515625,
      "rewards/final_brier_reward_step": 0.0395749993622303,
      "rewards/format_reward_step": 0.109375,
      "step": 12
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.5915297634071774,
      "aux_distill/mean_u": 0.24548161004521343,
      "aux_distill/n_active_tok": 61.77777777777778,
      "calib/answer_extract_rate": 0.1328125,
      "calib/auroc": 0.7301587301587302,
      "calib/avg_num_step_conf": 0.54296875,
      "calib/ece": 0.7495833333333334,
      "calib/final_conf_rate": 0.09375,
      "calib/format_rate": 0.078125,
      "calib/frac_conf_gt_0.9": 0.7083333333333334,
      "calib/gap": 0.08238095238095222,
      "calib/mean_conf": 0.8745833333333334,
      "calib/mu_c": 0.9466666666666667,
      "calib/mu_w": 0.8642857142857144,
      "calib/nonempty_final_conf_rate": 0.09375,
      "calib/nonempty_reasoning_rate": 0.16796875,
      "calib/nonempty_step_conf_rate": 0.125,
      "calib/pce": 0.7495833333333334,
      "calib/std_conf": 0.19004339782153853,
      "calib/step_conf_rate": 0.125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.08984375,
      "completions/max_length": 3045.0,
      "completions/max_terminated_length": 3045.0,
      "completions/mean_length": 630.265625,
      "completions/mean_terminated_length": 692.480712890625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.013866666666666666,
      "grad_norm": 0.008549676276743412,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.069,
      "num_tokens": 3528178.0,
      "reward": 0.05812598019838333,
      "reward_std": 0.10661930590867996,
      "rewards/accuracy_reward_step": 0.01171875,
      "rewards/final_brier_reward_step": 0.026408202946186066,
      "rewards/format_reward_step": 0.078125,
      "step": 13
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5545808651617595,
      "aux_distill/mean_u": 0.287806905984641,
      "aux_distill/n_active_tok": 61.285714285714285,
      "calib/answer_extract_rate": 0.234375,
      "calib/auroc": 0.6428571428571429,
      "calib/avg_num_step_conf": 0.87109375,
      "calib/ece": 0.6372045454545454,
      "calib/final_conf_rate": 0.171875,
      "calib/format_rate": 0.1328125,
      "calib/frac_conf_gt_0.9": 0.5,
      "calib/gap": 0.15560231660231671,
      "calib/mean_conf": 0.7962954545454545,
      "calib/mu_c": 0.9271428571428572,
      "calib/mu_w": 0.7715405405405404,
      "calib/nonempty_final_conf_rate": 0.171875,
      "calib/nonempty_reasoning_rate": 0.296875,
      "calib/nonempty_step_conf_rate": 0.20703125,
      "calib/pce": 0.6372045454545454,
      "calib/std_conf": 0.2955989743143614,
      "calib/step_conf_rate": 0.20703125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2701.0,
      "completions/max_terminated_length": 2701.0,
      "completions/mean_length": 581.75390625,
      "completions/mean_terminated_length": 615.4090576171875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.014933333333333333,
      "grad_norm": 0.010811959393322468,
      "learning_rate": 3.5e-06,
      "loss": 0.1161,
      "num_tokens": 3782507.0,
      "reward": 0.10872349888086319,
      "reward_std": 0.18449640274047852,
      "rewards/accuracy_reward_step": 0.02734375,
      "rewards/final_brier_reward_step": 0.05729074776172638,
      "rewards/format_reward_step": 0.1328125,
      "step": 14
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.534116182861657,
      "aux_distill/mean_u": 0.3509062857245206,
      "aux_distill/n_active_tok": 85.51724137931035,
      "calib/answer_extract_rate": 0.2109375,
      "calib/auroc": 0.6041666666666666,
      "calib/avg_num_step_conf": 1.20703125,
      "calib/ece": 0.655909090909091,
      "calib/final_conf_rate": 0.171875,
      "calib/format_rate": 0.1484375,
      "calib/frac_conf_gt_0.9": 0.5454545454545454,
      "calib/gap": 0.125,
      "calib/mean_conf": 0.8377272727272728,
      "calib/mu_c": 0.94,
      "calib/mu_w": 0.815,
      "calib/nonempty_final_conf_rate": 0.171875,
      "calib/nonempty_reasoning_rate": 0.28515625,
      "calib/nonempty_step_conf_rate": 0.23828125,
      "calib/pce": 0.655909090909091,
      "calib/std_conf": 0.21094876194301132,
      "calib/step_conf_rate": 0.23828125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3063.0,
      "completions/max_terminated_length": 3063.0,
      "completions/mean_length": 663.78125,
      "completions/mean_terminated_length": 710.9957885742188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 3.0,
      "epoch": 0.016,
      "grad_norm": 0.009891645051538944,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.1809,
      "num_tokens": 4060315.0,
      "reward": 0.12209217995405197,
      "reward_std": 0.2036409080028534,
      "rewards/accuracy_reward_step": 0.03125,
      "rewards/final_brier_reward_step": 0.06449687480926514,
      "rewards/format_reward_step": 0.1484375,
      "step": 15
    },
    {
      "aux_distill/lambda": 0.10000000000000003,
      "aux_distill/loss": 0.551109317690134,
      "aux_distill/mean_u": 0.41506540252915974,
      "aux_distill/n_active_tok": 92.29166666666667,
      "calib/answer_extract_rate": 0.23828125,
      "calib/auroc": 0.5151515151515151,
      "calib/avg_num_step_conf": 1.04296875,
      "calib/ece": 0.6365238095238095,
      "calib/final_conf_rate": 0.1640625,
      "calib/format_rate": 0.140625,
      "calib/frac_conf_gt_0.9": 0.5714285714285714,
      "calib/gap": 0.013373737373737482,
      "calib/mean_conf": 0.8110476190476189,
      "calib/mu_c": 0.8215555555555556,
      "calib/mu_w": 0.8081818181818181,
      "calib/nonempty_final_conf_rate": 0.1640625,
      "calib/nonempty_reasoning_rate": 0.29296875,
      "calib/nonempty_step_conf_rate": 0.21484375,
      "calib/pce": 0.6166428571428573,
      "calib/std_conf": 0.2606996038483987,
      "calib/step_conf_rate": 0.21484375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0703125,
      "completions/max_length": 2850.0,
      "completions/max_terminated_length": 2850.0,
      "completions/mean_length": 599.62109375,
      "completions/mean_terminated_length": 644.9706420898438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.017066666666666667,
      "grad_norm": 0.009639227762818336,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.1095,
      "num_tokens": 4322666.0,
      "reward": 0.11668315529823303,
      "reward_std": 0.21993079781532288,
      "rewards/accuracy_reward_step": 0.0390625,
      "rewards/final_brier_reward_step": 0.05367880314588547,
      "rewards/format_reward_step": 0.140625,
      "step": 16
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5508362718166844,
      "aux_distill/mean_u": 0.3202794545006914,
      "aux_distill/n_active_tok": 119.35483870967742,
      "calib/answer_extract_rate": 0.359375,
      "calib/auroc": 0.5582089552238806,
      "calib/avg_num_step_conf": 1.7890625,
      "calib/ece": 0.5943902439024391,
      "calib/final_conf_rate": 0.3203125,
      "calib/format_rate": 0.26171875,
      "calib/frac_conf_gt_0.9": 0.5,
      "calib/gap": 0.041472636815920505,
      "calib/mean_conf": 0.748780487804878,
      "calib/mu_c": 0.7826666666666667,
      "calib/mu_w": 0.7411940298507462,
      "calib/nonempty_final_conf_rate": 0.3203125,
      "calib/nonempty_reasoning_rate": 0.42578125,
      "calib/nonempty_step_conf_rate": 0.3515625,
      "calib/pce": 0.5801219512195123,
      "calib/std_conf": 0.3001153853059237,
      "calib/step_conf_rate": 0.3515625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2843.0,
      "completions/max_terminated_length": 2843.0,
      "completions/mean_length": 535.5625,
      "completions/mean_terminated_length": 555.0769653320312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 4.0,
      "epoch": 0.018133333333333335,
      "grad_norm": 0.012681329622864723,
      "learning_rate": 4.25e-06,
      "loss": 0.1619,
      "num_tokens": 4563298.0,
      "reward": 0.22102656960487366,
      "reward_std": 0.3276210427284241,
      "rewards/accuracy_reward_step": 0.0625,
      "rewards/final_brier_reward_step": 0.11783437430858612,
      "rewards/format_reward_step": 0.26171875,
      "step": 17
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5232953188880798,
      "aux_distill/mean_u": 0.3262361357347477,
      "aux_distill/n_active_tok": 98.48387096774194,
      "calib/answer_extract_rate": 0.33203125,
      "calib/auroc": 0.6807628524046434,
      "calib/avg_num_step_conf": 1.47265625,
      "calib/ece": 0.5537157894736842,
      "calib/final_conf_rate": 0.296875,
      "calib/format_rate": 0.20703125,
      "calib/frac_conf_gt_0.9": 0.39473684210526316,
      "calib/gap": 0.20301558872305125,
      "calib/mean_conf": 0.6721368421052633,
      "calib/mu_c": 0.8511111111111112,
      "calib/mu_w": 0.6480955223880599,
      "calib/nonempty_final_conf_rate": 0.296875,
      "calib/nonempty_reasoning_rate": 0.4140625,
      "calib/nonempty_step_conf_rate": 0.328125,
      "calib/pce": 0.5537157894736842,
      "calib/std_conf": 0.3439237510515901,
      "calib/step_conf_rate": 0.328125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2970.0,
      "completions/max_terminated_length": 2970.0,
      "completions/mean_length": 554.59765625,
      "completions/mean_terminated_length": 572.4879150390625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 5.0,
      "epoch": 0.0192,
      "grad_norm": 0.009989089332520962,
      "learning_rate": 4.5e-06,
      "loss": 0.1477,
      "num_tokens": 4815995.0,
      "reward": 0.17580628395080566,
      "reward_std": 0.2827049195766449,
      "rewards/accuracy_reward_step": 0.03515625,
      "rewards/final_brier_reward_step": 0.10942507535219193,
      "rewards/format_reward_step": 0.20703125,
      "step": 18
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5288848485797644,
      "aux_distill/mean_u": 0.3808616730169923,
      "aux_distill/n_active_tok": 209.78125,
      "calib/answer_extract_rate": 0.62109375,
      "calib/auroc": 0.5075464396284829,
      "calib/avg_num_step_conf": 3.19140625,
      "calib/ece": 0.45987161290322576,
      "calib/final_conf_rate": 0.60546875,
      "calib/format_rate": 0.4921875,
      "calib/frac_conf_gt_0.9": 0.24516129032258063,
      "calib/gap": 0.028739899380805012,
      "calib/mean_conf": 0.5408251612903225,
      "calib/mu_c": 0.5660421052631579,
      "calib/mu_w": 0.5373022058823529,
      "calib/nonempty_final_conf_rate": 0.60546875,
      "calib/nonempty_reasoning_rate": 0.76953125,
      "calib/nonempty_step_conf_rate": 0.671875,
      "calib/pce": 0.43905806451612905,
      "calib/std_conf": 0.3514235209036652,
      "calib/step_conf_rate": 0.671875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3024.0,
      "completions/max_terminated_length": 3024.0,
      "completions/mean_length": 405.7265625,
      "completions/mean_terminated_length": 420.5101318359375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 22.0,
      "epoch": 0.020266666666666665,
      "grad_norm": 0.015777073800563812,
      "learning_rate": 4.75e-06,
      "loss": 0.1692,
      "num_tokens": 5024621.0,
      "reward": 0.4331209063529968,
      "reward_std": 0.4471513032913208,
      "rewards/accuracy_reward_step": 0.078125,
      "rewards/final_brier_reward_step": 0.29592931270599365,
      "rewards/format_reward_step": 0.4921875,
      "step": 19
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5358412060886621,
      "aux_distill/mean_u": 0.3330040179088781,
      "aux_distill/n_active_tok": 235.40625,
      "calib/answer_extract_rate": 0.7734375,
      "calib/auroc": 0.5045928281222398,
      "calib/avg_num_step_conf": 3.58203125,
      "calib/ece": 0.3642105263157895,
      "calib/final_conf_rate": 0.7421875,
      "calib/format_rate": 0.640625,
      "calib/frac_conf_gt_0.9": 0.15789473684210525,
      "calib/gap": 0.015520226108461432,
      "calib/mean_conf": 0.46642105263157896,
      "calib/mu_c": 0.47891891891891897,
      "calib/mu_w": 0.46339869281045754,
      "calib/nonempty_final_conf_rate": 0.7421875,
      "calib/nonempty_reasoning_rate": 0.90625,
      "calib/nonempty_step_conf_rate": 0.83203125,
      "calib/pce": 0.31794736842105265,
      "calib/std_conf": 0.3506382631099845,
      "calib/step_conf_rate": 0.83203125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01953125,
      "completions/max_length": 2922.0,
      "completions/max_terminated_length": 2922.0,
      "completions/mean_length": 315.515625,
      "completions/mean_terminated_length": 321.8008117675781,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 15.0,
      "epoch": 0.021333333333333333,
      "grad_norm": 0.015699947252869606,
      "learning_rate": 5e-06,
      "loss": 0.1381,
      "num_tokens": 5210265.0,
      "reward": 0.6040115356445312,
      "reward_std": 0.46912193298339844,
      "rewards/accuracy_reward_step": 0.15234375,
      "rewards/final_brier_reward_step": 0.4150542914867401,
      "rewards/format_reward_step": 0.640625,
      "step": 20
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5463973861187696,
      "aux_distill/mean_u": 0.32214415168677285,
      "aux_distill/n_active_tok": 264.90625,
      "calib/answer_extract_rate": 0.828125,
      "calib/auroc": 0.4245585874799358,
      "calib/avg_num_step_conf": 3.9140625,
      "calib/ece": 0.319245145631068,
      "calib/final_conf_rate": 0.8046875,
      "calib/format_rate": 0.75,
      "calib/frac_conf_gt_0.9": 0.10679611650485436,
      "calib/gap": -0.08127889245585873,
      "calib/mean_conf": 0.3720169902912622,
      "calib/mu_c": 0.30178571428571427,
      "calib/mu_w": 0.383064606741573,
      "calib/nonempty_final_conf_rate": 0.8046875,
      "calib/nonempty_reasoning_rate": 0.9453125,
      "calib/nonempty_step_conf_rate": 0.89453125,
      "calib/pce": 0.2776699029126214,
      "calib/std_conf": 0.3370346662655462,
      "calib/step_conf_rate": 0.89453125,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2032.0,
      "completions/max_terminated_length": 2032.0,
      "completions/mean_length": 318.78515625,
      "completions/mean_terminated_length": 321.2952880859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 25.0,
      "epoch": 0.0224,
      "grad_norm": 0.016127225011587143,
      "learning_rate": 4.9722222222222224e-06,
      "loss": 0.2055,
      "num_tokens": 5394834.0,
      "reward": 0.6895323991775513,
      "reward_std": 0.4213147759437561,
      "rewards/accuracy_reward_step": 0.11328125,
      "rewards/final_brier_reward_step": 0.5157834887504578,
      "rewards/format_reward_step": 0.75,
      "step": 21
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.500660709105432,
      "aux_distill/mean_u": 0.3673933193010618,
      "aux_distill/n_active_tok": 297.125,
      "calib/answer_extract_rate": 0.890625,
      "calib/auroc": 0.39062499999999994,
      "calib/avg_num_step_conf": 4.60546875,
      "calib/ece": 0.3734529680365296,
      "calib/final_conf_rate": 0.85546875,
      "calib/format_rate": 0.8203125,
      "calib/frac_conf_gt_0.9": 0.1689497716894977,
      "calib/gap": -0.13653755787037036,
      "calib/mean_conf": 0.4323004566210046,
      "calib/mu_c": 0.3125962962962963,
      "calib/mu_w": 0.4491338541666667,
      "calib/nonempty_final_conf_rate": 0.85546875,
      "calib/nonempty_reasoning_rate": 0.9765625,
      "calib/nonempty_step_conf_rate": 0.94921875,
      "calib/pce": 0.3412328767123287,
      "calib/std_conf": 0.34979617502728416,
      "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": 2701.0,
      "completions/max_terminated_length": 2701.0,
      "completions/mean_length": 326.50390625,
      "completions/mean_terminated_length": 327.7843322753906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 48.0,
      "epoch": 0.023466666666666667,
      "grad_norm": 0.012950566597282887,
      "learning_rate": 4.944444444444445e-06,
      "loss": 0.2034,
      "num_tokens": 5580235.0,
      "reward": 0.727470338344574,
      "reward_std": 0.36542022228240967,
      "rewards/accuracy_reward_step": 0.10546875,
      "rewards/final_brier_reward_step": 0.529159426689148,
      "rewards/format_reward_step": 0.8203125,
      "step": 22
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5794403096660972,
      "aux_distill/mean_u": 0.3093092033622721,
      "aux_distill/n_active_tok": 255.09375,
      "calib/answer_extract_rate": 0.8203125,
      "calib/auroc": 0.4833436788744051,
      "calib/avg_num_step_conf": 3.78515625,
      "calib/ece": 0.2792660194174757,
      "calib/final_conf_rate": 0.8046875,
      "calib/format_rate": 0.75390625,
      "calib/frac_conf_gt_0.9": 0.07766990291262135,
      "calib/gap": -0.027863107800538023,
      "calib/mean_conf": 0.3438407766990291,
      "calib/mu_c": 0.3196296296296296,
      "calib/mu_w": 0.3474927374301676,
      "calib/nonempty_final_conf_rate": 0.8046875,
      "calib/nonempty_reasoning_rate": 0.94921875,
      "calib/nonempty_step_conf_rate": 0.9140625,
      "calib/pce": 0.24601941747572814,
      "calib/std_conf": 0.3269325359987017,
      "calib/step_conf_rate": 0.9140625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2109.0,
      "completions/max_terminated_length": 2109.0,
      "completions/mean_length": 303.39453125,
      "completions/mean_terminated_length": 305.7834777832031,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 2.0,
      "epoch": 0.024533333333333334,
      "grad_norm": 0.01320746261626482,
      "learning_rate": 4.9166666666666665e-06,
      "loss": 0.2314,
      "num_tokens": 5761840.0,
      "reward": 0.7033728361129761,
      "reward_std": 0.4077947735786438,
      "rewards/accuracy_reward_step": 0.10546875,
      "rewards/final_brier_reward_step": 0.5473706722259521,
      "rewards/format_reward_step": 0.75390625,
      "step": 23
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5133433574810624,
      "aux_distill/mean_u": 0.3777117190213353,
      "aux_distill/n_active_tok": 289.9375,
      "calib/answer_extract_rate": 0.91796875,
      "calib/auroc": 0.5641506602641055,
      "calib/avg_num_step_conf": 4.4375,
      "calib/ece": 0.2902073913043478,
      "calib/final_conf_rate": 0.8984375,
      "calib/format_rate": 0.84765625,
      "calib/frac_conf_gt_0.9": 0.1,
      "calib/gap": 0.06962457983193282,
      "calib/mean_conf": 0.3789030434782609,
      "calib/mu_c": 0.43823529411764706,
      "calib/mu_w": 0.36861071428571424,
      "calib/nonempty_final_conf_rate": 0.8984375,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.94921875,
      "calib/pce": 0.26064217391304345,
      "calib/std_conf": 0.3273176261401784,
      "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": 2679.0,
      "completions/max_terminated_length": 2679.0,
      "completions/mean_length": 305.03125,
      "completions/mean_terminated_length": 306.22747802734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 6.0,
      "epoch": 0.0256,
      "grad_norm": 0.01090239454060793,
      "learning_rate": 4.888888888888889e-06,
      "loss": 0.1741,
      "num_tokens": 5944440.0,
      "reward": 0.802044153213501,
      "reward_std": 0.35184115171432495,
      "rewards/accuracy_reward_step": 0.1328125,
      "rewards/final_brier_reward_step": 0.623619556427002,
      "rewards/format_reward_step": 0.84765625,
      "step": 24
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5091779874637723,
      "aux_distill/mean_u": 0.3399544439281606,
      "aux_distill/n_active_tok": 275.5,
      "calib/answer_extract_rate": 0.93359375,
      "calib/auroc": 0.6165458141067898,
      "calib/avg_num_step_conf": 4.28515625,
      "calib/ece": 0.220200826446281,
      "calib/final_conf_rate": 0.9453125,
      "calib/format_rate": 0.90234375,
      "calib/frac_conf_gt_0.9": 0.07024793388429752,
      "calib/gap": 0.10513278839815426,
      "calib/mean_conf": 0.3250223140495867,
      "calib/mu_c": 0.4140810810810811,
      "calib/mu_w": 0.30894829268292684,
      "calib/nonempty_final_conf_rate": 0.9453125,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.19616528925619836,
      "calib/std_conf": 0.3018464867580852,
      "calib/step_conf_rate": 0.98046875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 2797.0,
      "completions/max_terminated_length": 2797.0,
      "completions/mean_length": 254.38671875,
      "completions/mean_terminated_length": 256.3897705078125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.02666666666666667,
      "grad_norm": 0.0111892344430089,
      "learning_rate": 4.861111111111111e-06,
      "loss": 0.1203,
      "num_tokens": 6112787.0,
      "reward": 0.8762689828872681,
      "reward_std": 0.3159981966018677,
      "rewards/accuracy_reward_step": 0.14453125,
      "rewards/final_brier_reward_step": 0.7056629061698914,
      "rewards/format_reward_step": 0.90234375,
      "step": 25
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5155609296634793,
      "aux_distill/mean_u": 0.4098851851699954,
      "aux_distill/n_active_tok": 251.9375,
      "calib/answer_extract_rate": 0.9453125,
      "calib/auroc": 0.4148527528809219,
      "calib/avg_num_step_conf": 3.87890625,
      "calib/ece": 0.2528608510638298,
      "calib/final_conf_rate": 0.91796875,
      "calib/format_rate": 0.90625,
      "calib/frac_conf_gt_0.9": 0.05531914893617021,
      "calib/gap": -0.08365330772513868,
      "calib/mean_conf": 0.2944582978723404,
      "calib/mu_c": 0.21863636363636363,
      "calib/mu_w": 0.3022896713615023,
      "calib/nonempty_final_conf_rate": 0.91796875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.22685106382978726,
      "calib/std_conf": 0.3056578070766229,
      "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": 2999.0,
      "completions/max_terminated_length": 2999.0,
      "completions/mean_length": 243.84765625,
      "completions/mean_terminated_length": 243.84765625,
      "completions/min_length": 23.0,
      "completions/min_terminated_length": 23.0,
      "epoch": 0.027733333333333332,
      "grad_norm": 0.011879602447152138,
      "learning_rate": 4.833333333333333e-06,
      "loss": 0.1231,
      "num_tokens": 6280452.0,
      "reward": 0.8439472913742065,
      "reward_std": 0.28856855630874634,
      "rewards/accuracy_reward_step": 0.0859375,
      "rewards/final_brier_reward_step": 0.6957069635391235,
      "rewards/format_reward_step": 0.90625,
      "step": 26
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4900315869599581,
      "aux_distill/mean_u": 0.3588746115217724,
      "aux_distill/n_active_tok": 251.34375,
      "calib/answer_extract_rate": 0.94921875,
      "calib/auroc": 0.5614362806143629,
      "calib/avg_num_step_conf": 3.9296875,
      "calib/ece": 0.2168344398340249,
      "calib/final_conf_rate": 0.94140625,
      "calib/format_rate": 0.92578125,
      "calib/frac_conf_gt_0.9": 0.03734439834024896,
      "calib/gap": 0.07003034454130347,
      "calib/mean_conf": 0.2768170124481327,
      "calib/mu_c": 0.34045454545454545,
      "calib/mu_w": 0.270424200913242,
      "calib/nonempty_final_conf_rate": 0.94140625,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.9765625,
      "calib/pce": 0.20118257261410788,
      "calib/std_conf": 0.27174081454851323,
      "calib/step_conf_rate": 0.9765625,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 1898.0,
      "completions/max_terminated_length": 1898.0,
      "completions/mean_length": 226.91015625,
      "completions/mean_terminated_length": 228.6968536376953,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 11.0,
      "epoch": 0.0288,
      "grad_norm": 0.01075553335249424,
      "learning_rate": 4.805555555555556e-06,
      "loss": 0.0758,
      "num_tokens": 6443757.0,
      "reward": 0.8853782415390015,
      "reward_std": 0.24783457815647125,
      "rewards/accuracy_reward_step": 0.0859375,
      "rewards/final_brier_reward_step": 0.7590377330780029,
      "rewards/format_reward_step": 0.92578125,
      "step": 27
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5010905349627137,
      "aux_distill/mean_u": 0.35061370361301236,
      "aux_distill/n_active_tok": 253.0625,
      "calib/answer_extract_rate": 0.953125,
      "calib/auroc": 0.40372168284789645,
      "calib/avg_num_step_conf": 3.90625,
      "calib/ece": 0.23082326530612246,
      "calib/final_conf_rate": 0.95703125,
      "calib/format_rate": 0.92578125,
      "calib/frac_conf_gt_0.9": 0.0326530612244898,
      "calib/gap": -0.06636441374159821,
      "calib/mean_conf": 0.2655848979591837,
      "calib/mu_c": 0.20978461538461537,
      "calib/mu_w": 0.2761490291262136,
      "calib/nonempty_final_conf_rate": 0.95703125,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.96875,
      "calib/pce": 0.1686122448979592,
      "calib/std_conf": 0.2550969035094147,
      "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": 2758.0,
      "completions/max_terminated_length": 2758.0,
      "completions/mean_length": 226.02734375,
      "completions/mean_terminated_length": 226.02734375,
      "completions/min_length": 25.0,
      "completions/min_terminated_length": 25.0,
      "epoch": 0.029866666666666666,
      "grad_norm": 0.010541618801653385,
      "learning_rate": 4.777777777777778e-06,
      "loss": 0.0724,
      "num_tokens": 6608564.0,
      "reward": 0.8957502245903015,
      "reward_std": 0.2513200044631958,
      "rewards/accuracy_reward_step": 0.15234375,
      "rewards/final_brier_reward_step": 0.7133753895759583,
      "rewards/format_reward_step": 0.92578125,
      "step": 28
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5015281261876225,
      "aux_distill/mean_u": 0.37213500914810527,
      "aux_distill/n_active_tok": 242.53125,
      "calib/answer_extract_rate": 0.953125,
      "calib/auroc": 0.5502812939521801,
      "calib/avg_num_step_conf": 3.734375,
      "calib/ece": 0.21642570281124499,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.94140625,
      "calib/frac_conf_gt_0.9": 0.028112449799196786,
      "calib/gap": 0.0666518987341772,
      "calib/mean_conf": 0.2557269076305221,
      "calib/mu_c": 0.31916666666666665,
      "calib/mu_w": 0.25251476793248945,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.21197991967871488,
      "calib/std_conf": 0.24414579528695696,
      "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": 1377.0,
      "completions/max_terminated_length": 1377.0,
      "completions/mean_length": 211.65625,
      "completions/mean_terminated_length": 212.4862823486328,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 44.0,
      "epoch": 0.030933333333333334,
      "grad_norm": 0.011848612688481808,
      "learning_rate": 4.75e-06,
      "loss": 0.0688,
      "num_tokens": 6769876.0,
      "reward": 0.8962675333023071,
      "reward_std": 0.19631339609622955,
      "rewards/accuracy_reward_step": 0.046875,
      "rewards/final_brier_reward_step": 0.8042538166046143,
      "rewards/format_reward_step": 0.94140625,
      "step": 29
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48358502611517906,
      "aux_distill/mean_u": 0.31133066449118657,
      "aux_distill/n_active_tok": 217.15625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.47853773584905657,
      "calib/avg_num_step_conf": 3.36328125,
      "calib/ece": 0.17899841269841268,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.023809523809523808,
      "calib/gap": -0.0208839622641509,
      "calib/mean_conf": 0.2563190476190476,
      "calib/mu_c": 0.23875000000000002,
      "calib/mu_w": 0.2596339622641509,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.13829365079365077,
      "calib/std_conf": 0.23551776218335396,
      "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": 627.0,
      "completions/max_terminated_length": 627.0,
      "completions/mean_length": 172.63671875,
      "completions/mean_terminated_length": 173.31373596191406,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 45.0,
      "epoch": 0.032,
      "grad_norm": 0.01241193525493145,
      "learning_rate": 4.722222222222222e-06,
      "loss": 0.0386,
      "num_tokens": 6921055.0,
      "reward": 0.9601128101348877,
      "reward_std": 0.15688207745552063,
      "rewards/accuracy_reward_step": 0.16015625,
      "rewards/final_brier_reward_step": 0.7796005010604858,
      "rewards/format_reward_step": 0.98046875,
      "step": 30
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.49283486418426037,
      "aux_distill/mean_u": 0.31671603956800987,
      "aux_distill/n_active_tok": 202.46875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5469432314410481,
      "calib/avg_num_step_conf": 3.1171875,
      "calib/ece": 0.18115298804780877,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.9609375,
      "calib/frac_conf_gt_0.9": 0.01593625498007968,
      "calib/gap": 0.05664088924176261,
      "calib/mean_conf": 0.2451418326693227,
      "calib/mu_c": 0.2968181818181818,
      "calib/mu_w": 0.2401772925764192,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.16932270916334663,
      "calib/std_conf": 0.22780424523358234,
      "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": 460.0,
      "completions/max_terminated_length": 460.0,
      "completions/mean_length": 159.74609375,
      "completions/mean_terminated_length": 160.37255859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 49.0,
      "epoch": 0.03306666666666667,
      "grad_norm": 0.012938912026584148,
      "learning_rate": 4.694444444444445e-06,
      "loss": 0.0724,
      "num_tokens": 7067862.0,
      "reward": 0.9331058263778687,
      "reward_std": 0.1786496937274933,
      "rewards/accuracy_reward_step": 0.0859375,
      "rewards/final_brier_reward_step": 0.8193366527557373,
      "rewards/format_reward_step": 0.9609375,
      "step": 31
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5122799929231405,
      "aux_distill/mean_u": 0.3455168102658182,
      "aux_distill/n_active_tok": 202.125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5427272727272726,
      "calib/avg_num_step_conf": 3.1328125,
      "calib/ece": 0.17164000000000001,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.016,
      "calib/gap": 0.05277272727272722,
      "calib/mean_conf": 0.24356,
      "calib/mu_c": 0.29,
      "calib/mu_w": 0.23722727272727276,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.14759999999999998,
      "calib/std_conf": 0.2167369982259605,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 483.0,
      "completions/max_terminated_length": 483.0,
      "completions/mean_length": 157.83203125,
      "completions/mean_terminated_length": 158.45098876953125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 43.0,
      "epoch": 0.034133333333333335,
      "grad_norm": 0.013937647454440594,
      "learning_rate": 4.666666666666667e-06,
      "loss": 0.0398,
      "num_tokens": 7214971.0,
      "reward": 0.9550351500511169,
      "reward_std": 0.15868309140205383,
      "rewards/accuracy_reward_step": 0.1171875,
      "rewards/final_brier_reward_step": 0.8202266097068787,
      "rewards/format_reward_step": 0.97265625,
      "step": 32
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5047999192029238,
      "aux_distill/mean_u": 0.3599142013757496,
      "aux_distill/n_active_tok": 199.25,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5623517786561265,
      "calib/avg_num_step_conf": 3.125,
      "calib/ece": 0.13408730158730156,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": 0.034723320158102755,
      "calib/mean_conf": 0.19694444444444445,
      "calib/mu_c": 0.22863636363636364,
      "calib/mu_w": 0.19391304347826088,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.12186507936507934,
      "calib/std_conf": 0.19079145123920305,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1433.0,
      "completions/max_terminated_length": 1433.0,
      "completions/mean_length": 167.23828125,
      "completions/mean_terminated_length": 167.23828125,
      "completions/min_length": 49.0,
      "completions/min_terminated_length": 49.0,
      "epoch": 0.0352,
      "grad_norm": 0.013600134290754795,
      "learning_rate": 4.638888888888889e-06,
      "loss": 0.0804,
      "num_tokens": 7364656.0,
      "reward": 0.9592347145080566,
      "reward_std": 0.12692934274673462,
      "rewards/accuracy_reward_step": 0.0859375,
      "rewards/final_brier_reward_step": 0.8559695482254028,
      "rewards/format_reward_step": 0.9765625,
      "step": 33
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47679288033396006,
      "aux_distill/mean_u": 0.33165973203230903,
      "aux_distill/n_active_tok": 215.75,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.5446511627906977,
      "calib/avg_num_step_conf": 3.375,
      "calib/ece": 0.11831999999999998,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.01176079734219268,
      "calib/mean_conf": 0.18560000000000001,
      "calib/mu_c": 0.1957142857142857,
      "calib/mu_w": 0.18395348837209302,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.08195999999999999,
      "calib/std_conf": 0.17300127167162674,
      "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": 3003.0,
      "completions/max_terminated_length": 3003.0,
      "completions/mean_length": 180.69921875,
      "completions/mean_terminated_length": 180.69921875,
      "completions/min_length": 14.0,
      "completions/min_terminated_length": 14.0,
      "epoch": 0.03626666666666667,
      "grad_norm": 0.01341802254319191,
      "learning_rate": 4.611111111111112e-06,
      "loss": 0.0663,
      "num_tokens": 7516027.0,
      "reward": 0.9679831862449646,
      "reward_std": 0.1320965737104416,
      "rewards/accuracy_reward_step": 0.13671875,
      "rewards/final_brier_reward_step": 0.826591432094574,
      "rewards/format_reward_step": 0.97265625,
      "step": 34
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5014631142839789,
      "aux_distill/mean_u": 0.3500040735053896,
      "aux_distill/n_active_tok": 245.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5164835164835164,
      "calib/avg_num_step_conf": 3.859375,
      "calib/ece": 0.12572549019607843,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.00392156862745098,
      "calib/gap": 0.025903540903540884,
      "calib/mean_conf": 0.13670588235294118,
      "calib/mu_c": 0.16047619047619047,
      "calib/mu_w": 0.13457264957264958,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0900392156862745,
      "calib/std_conf": 0.15651139081739068,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 488.0,
      "completions/max_terminated_length": 488.0,
      "completions/mean_length": 173.75390625,
      "completions/mean_terminated_length": 174.435302734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 55.0,
      "epoch": 0.037333333333333336,
      "grad_norm": 0.012046094983816147,
      "learning_rate": 4.583333333333333e-06,
      "loss": 0.0359,
      "num_tokens": 7669764.0,
      "reward": 0.9877500534057617,
      "reward_std": 0.06010032445192337,
      "rewards/accuracy_reward_step": 0.08203125,
      "rewards/final_brier_reward_step": 0.8973749876022339,
      "rewards/format_reward_step": 0.99609375,
      "step": 35
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4936283128336072,
      "aux_distill/mean_u": 0.33273859315272475,
      "aux_distill/n_active_tok": 251.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5374149659863946,
      "calib/avg_num_step_conf": 3.94921875,
      "calib/ece": 0.15101562500000001,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.01018436360051267,
      "calib/mean_conf": 0.10523437500000002,
      "calib/mu_c": 0.11346938775510206,
      "calib/mu_w": 0.10328502415458939,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.032421874999999996,
      "calib/std_conf": 0.12121755264135377,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 448.0,
      "completions/max_terminated_length": 448.0,
      "completions/mean_length": 178.6484375,
      "completions/mean_terminated_length": 179.34902954101562,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 49.0,
      "epoch": 0.0384,
      "grad_norm": 0.011449523270130157,
      "learning_rate": 4.555555555555556e-06,
      "loss": 0.0791,
      "num_tokens": 7818210.0,
      "reward": 1.0088348388671875,
      "reward_std": 0.05037228763103485,
      "rewards/accuracy_reward_step": 0.19140625,
      "rewards/final_brier_reward_step": 0.8262632489204407,
      "rewards/format_reward_step": 1.0,
      "step": 36
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4734587259590626,
      "aux_distill/mean_u": 0.29364758001900115,
      "aux_distill/n_active_tok": 254.03125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5375,
      "calib/avg_num_step_conf": 3.97265625,
      "calib/ece": 0.08688976377952756,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0009583333333333666,
      "calib/mean_conf": 0.08051181102362205,
      "calib/mu_c": 0.07966666666666665,
      "calib/mu_w": 0.08062500000000002,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.02464566929133858,
      "calib/std_conf": 0.10341479555286022,
      "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": 788.0,
      "completions/max_terminated_length": 788.0,
      "completions/mean_length": 185.39453125,
      "completions/mean_terminated_length": 186.12158203125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 40.0,
      "epoch": 0.039466666666666664,
      "grad_norm": 0.012011139653623104,
      "learning_rate": 4.527777777777778e-06,
      "loss": 0.0557,
      "num_tokens": 7972767.0,
      "reward": 0.9930021166801453,
      "reward_std": 0.04802718758583069,
      "rewards/accuracy_reward_step": 0.1171875,
      "rewards/final_brier_reward_step": 0.8766292333602905,
      "rewards/format_reward_step": 0.9921875,
      "step": 37
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.5167172281071544,
      "aux_distill/mean_u": 0.35844834511413254,
      "aux_distill/n_active_tok": 270.875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5392903087478559,
      "calib/avg_num_step_conf": 4.2265625,
      "calib/ece": 0.12679687499999998,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.02404802744425384,
      "calib/mean_conf": 0.06781250000000001,
      "calib/mu_c": 0.08772727272727271,
      "calib/mu_w": 0.06367924528301887,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.011367187500000002,
      "calib/std_conf": 0.09589575639072878,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 421.0,
      "completions/max_terminated_length": 421.0,
      "completions/mean_length": 186.63671875,
      "completions/mean_terminated_length": 187.36863708496094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 64.0,
      "epoch": 0.04053333333333333,
      "grad_norm": 0.011758743785321712,
      "learning_rate": 4.5e-06,
      "loss": 0.0434,
      "num_tokens": 8127434.0,
      "reward": 1.0042815208435059,
      "reward_std": 0.03750614821910858,
      "rewards/accuracy_reward_step": 0.171875,
      "rewards/final_brier_reward_step": 0.8405945301055908,
      "rewards/format_reward_step": 0.99609375,
      "step": 38
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4876714339479804,
      "aux_distill/mean_u": 0.30279994241838204,
      "aux_distill/n_active_tok": 263.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.41560473954840155,
      "calib/avg_num_step_conf": 4.13671875,
      "calib/ece": 0.14603921568627454,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.018843058350100608,
      "calib/mean_conf": 0.04454901960784314,
      "calib/mu_c": 0.02880952380952381,
      "calib/mu_w": 0.04765258215962442,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.012941176470588234,
      "calib/std_conf": 0.08763388943479686,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 556.0,
      "completions/max_terminated_length": 556.0,
      "completions/mean_length": 185.75,
      "completions/mean_terminated_length": 186.4784393310547,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 67.0,
      "epoch": 0.0416,
      "grad_norm": 0.011680186726152897,
      "learning_rate": 4.472222222222223e-06,
      "loss": 0.0806,
      "num_tokens": 8281074.0,
      "reward": 0.9979601502418518,
      "reward_std": 0.02412526123225689,
      "rewards/accuracy_reward_step": 0.16796875,
      "rewards/final_brier_reward_step": 0.8318578004837036,
      "rewards/format_reward_step": 0.99609375,
      "step": 39
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4841670235618949,
      "aux_distill/mean_u": 0.30947448492696666,
      "aux_distill/n_active_tok": 258.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.45479855879462827,
      "calib/avg_num_step_conf": 4.046875,
      "calib/ece": 0.15710937500000002,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.015516977836008292,
      "calib/mean_conf": 0.03546875000000001,
      "calib/mu_c": 0.022558139534883725,
      "calib/mu_w": 0.03807511737089202,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.012304687500000001,
      "calib/std_conf": 0.06960108852192974,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 494.0,
      "completions/max_terminated_length": 494.0,
      "completions/mean_length": 179.8046875,
      "completions/mean_terminated_length": 180.5098114013672,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 53.0,
      "epoch": 0.042666666666666665,
      "grad_norm": 0.012068229727447033,
      "learning_rate": 4.444444444444444e-06,
      "loss": 0.0551,
      "num_tokens": 8433864.0,
      "reward": 1.0007379055023193,
      "reward_std": 0.01349399983882904,
      "rewards/accuracy_reward_step": 0.16796875,
      "rewards/final_brier_reward_step": 0.8335070610046387,
      "rewards/format_reward_step": 1.0,
      "step": 40
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.49721950199455023,
      "aux_distill/mean_u": 0.3209028292175142,
      "aux_distill/n_active_tok": 253.8125,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.46456508135168956,
      "calib/avg_num_step_conf": 3.984375,
      "calib/ece": 0.24701093749999997,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 8.498122653316162e-05,
      "calib/mean_conf": 0.0186140625,
      "calib/mu_c": 0.01867647058823529,
      "calib/mu_w": 0.01859148936170213,
      "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": 0.025049272489357725,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 410.0,
      "completions/max_terminated_length": 410.0,
      "completions/mean_length": 171.5859375,
      "completions/mean_terminated_length": 172.2588348388672,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 69.0,
      "epoch": 0.04373333333333333,
      "grad_norm": 0.013531737960875034,
      "learning_rate": 4.416666666666667e-06,
      "loss": 0.0318,
      "num_tokens": 8585038.0,
      "reward": 1.0044739246368408,
      "reward_std": 0.00973083171993494,
      "rewards/accuracy_reward_step": 0.265625,
      "rewards/final_brier_reward_step": 0.7433229684829712,
      "rewards/format_reward_step": 1.0,
      "step": 41
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48902952019125223,
      "aux_distill/mean_u": 0.32963415527338386,
      "aux_distill/n_active_tok": 246.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5018560978272737,
      "calib/avg_num_step_conf": 3.85546875,
      "calib/ece": 0.1583203125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0009378753138988959,
      "calib/mean_conf": 0.0156640625,
      "calib/mu_c": 0.014883720930232559,
      "calib/mu_w": 0.015821596244131455,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0030078125,
      "calib/std_conf": 0.026449178077892963,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 405.0,
      "completions/max_terminated_length": 405.0,
      "completions/mean_length": 169.76953125,
      "completions/mean_terminated_length": 170.435302734375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 61.0,
      "epoch": 0.0448,
      "grad_norm": 0.01112829614430666,
      "learning_rate": 4.388888888888889e-06,
      "loss": 0.0484,
      "num_tokens": 8732867.0,
      "reward": 1.0020275115966797,
      "reward_std": 0.0062466030940413475,
      "rewards/accuracy_reward_step": 0.16796875,
      "rewards/final_brier_reward_step": 0.8360863327980042,
      "rewards/format_reward_step": 1.0,
      "step": 42
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4769623838365078,
      "aux_distill/mean_u": 0.2908125661926928,
      "aux_distill/n_active_tok": 244.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4851258581235698,
      "calib/avg_num_step_conf": 3.82421875,
      "calib/ece": 0.16768627450980395,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0005804035781152486,
      "calib/mean_conf": 0.014823529411764708,
      "calib/mu_c": 0.014347826086956523,
      "calib/mu_w": 0.014928229665071771,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0010588235294117648,
      "calib/std_conf": 0.01987875591084262,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 373.0,
      "completions/max_terminated_length": 373.0,
      "completions/mean_length": 172.5859375,
      "completions/mean_terminated_length": 173.26275634765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 74.0,
      "epoch": 0.04586666666666667,
      "grad_norm": 0.01158448588103056,
      "learning_rate": 4.361111111111112e-06,
      "loss": 0.0236,
      "num_tokens": 8882273.0,
      "reward": 0.9964126944541931,
      "reward_std": 0.02189866080880165,
      "rewards/accuracy_reward_step": 0.18359375,
      "rewards/final_brier_reward_step": 0.8170441389083862,
      "rewards/format_reward_step": 0.9921875,
      "step": 43
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48727248795330524,
      "aux_distill/mean_u": 0.32562953287294033,
      "aux_distill/n_active_tok": 270.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5547118785105055,
      "calib/avg_num_step_conf": 4.23828125,
      "calib/ece": 0.1688235294117647,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0017994591221135846,
      "calib/mean_conf": 0.011568627450980393,
      "calib/mu_c": 0.013043478260869568,
      "calib/mu_w": 0.011244019138755983,
      "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": 0.015539014658191224,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 438.0,
      "completions/max_terminated_length": 438.0,
      "completions/mean_length": 188.04296875,
      "completions/mean_terminated_length": 188.78041076660156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 68.0,
      "epoch": 0.046933333333333334,
      "grad_norm": 0.012537293136119843,
      "learning_rate": 4.333333333333334e-06,
      "loss": 0.0531,
      "num_tokens": 9036732.0,
      "reward": 0.9982506036758423,
      "reward_std": 0.015460444614291191,
      "rewards/accuracy_reward_step": 0.1796875,
      "rewards/final_brier_reward_step": 0.8207198977470398,
      "rewards/format_reward_step": 0.99609375,
      "step": 44
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4742131344974041,
      "aux_distill/mean_u": 0.26243291542283825,
      "aux_distill/n_active_tok": 262.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4895609756097561,
      "calib/avg_num_step_conf": 4.10546875,
      "calib/ece": 0.18729411764705883,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0012634146341463422,
      "calib/mean_conf": 0.008784313725490198,
      "calib/mu_c": 0.009800000000000001,
      "calib/mu_w": 0.00853658536585366,
      "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": 0.013478106758144088,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2400.0,
      "completions/max_terminated_length": 2400.0,
      "completions/mean_length": 201.03515625,
      "completions/mean_terminated_length": 201.03515625,
      "completions/min_length": 45.0,
      "completions/min_terminated_length": 45.0,
      "epoch": 0.048,
      "grad_norm": 0.009973783977329731,
      "learning_rate": 4.305555555555556e-06,
      "loss": 0.0858,
      "num_tokens": 9193245.0,
      "reward": 0.997878909111023,
      "reward_std": 0.015608172863721848,
      "rewards/accuracy_reward_step": 0.1953125,
      "rewards/final_brier_reward_step": 0.8043515682220459,
      "rewards/format_reward_step": 0.99609375,
      "step": 45
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4932613065466285,
      "aux_distill/mean_u": 0.3355272456696168,
      "aux_distill/n_active_tok": 263.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.43321371610845294,
      "calib/avg_num_step_conf": 4.11328125,
      "calib/ece": 0.2507421875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0007480063795853274,
      "calib/mean_conf": 0.0070703125,
      "calib/mu_c": 0.0065151515151515155,
      "calib/mu_w": 0.007263157894736843,
      "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": 0.010402586512610399,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 452.0,
      "completions/max_terminated_length": 452.0,
      "completions/mean_length": 179.2421875,
      "completions/mean_terminated_length": 179.9451141357422,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 60.0,
      "epoch": 0.04906666666666667,
      "grad_norm": 0.010921996086835861,
      "learning_rate": 4.277777777777778e-06,
      "loss": 0.0575,
      "num_tokens": 9343899.0,
      "reward": 1.0016005039215088,
      "reward_std": 0.0036760293878614902,
      "rewards/accuracy_reward_step": 0.2578125,
      "rewards/final_brier_reward_step": 0.7453886270523071,
      "rewards/format_reward_step": 1.0,
      "step": 46
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4907516334205866,
      "aux_distill/mean_u": 0.321547496121402,
      "aux_distill/n_active_tok": 279.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.45729638009049767,
      "calib/avg_num_step_conf": 4.37890625,
      "calib/ece": 0.1972265625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.003495475113122172,
      "calib/mean_conf": 0.0085546875,
      "calib/mu_c": 0.00576923076923077,
      "calib/mu_w": 0.009264705882352942,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.001328125,
      "calib/std_conf": 0.022943990428374566,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 462.0,
      "completions/max_terminated_length": 462.0,
      "completions/mean_length": 189.0625,
      "completions/mean_terminated_length": 189.80393981933594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 76.0,
      "epoch": 0.050133333333333335,
      "grad_norm": 0.011003137566149235,
      "learning_rate": 4.25e-06,
      "loss": 0.0596,
      "num_tokens": 9498275.0,
      "reward": 1.000872015953064,
      "reward_std": 0.00331785692833364,
      "rewards/accuracy_reward_step": 0.203125,
      "rewards/final_brier_reward_step": 0.7986191511154175,
      "rewards/format_reward_step": 1.0,
      "step": 47
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4879733482375741,
      "aux_distill/mean_u": 0.30240783007869915,
      "aux_distill/n_active_tok": 255.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.45755091245702195,
      "calib/avg_num_step_conf": 3.98828125,
      "calib/ece": 0.2168359375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0013938111610685014,
      "calib/mean_conf": 0.0058203125,
      "calib/mu_c": 0.004736842105263158,
      "calib/mu_w": 0.00613065326633166,
      "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": 0.008844961130629334,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 492.0,
      "completions/max_terminated_length": 492.0,
      "completions/mean_length": 181.05859375,
      "completions/mean_terminated_length": 181.7686309814453,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 60.0,
      "epoch": 0.0512,
      "grad_norm": 0.011607394553720951,
      "learning_rate": 4.222222222222223e-06,
      "loss": 0.0429,
      "num_tokens": 9648314.0,
      "reward": 1.0009984970092773,
      "reward_std": 0.002487377729266882,
      "rewards/accuracy_reward_step": 0.22265625,
      "rewards/final_brier_reward_step": 0.7793409824371338,
      "rewards/format_reward_step": 1.0,
      "step": 48
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.490774248726666,
      "aux_distill/mean_u": 0.32144850803263364,
      "aux_distill/n_active_tok": 261.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5138048780487804,
      "calib/avg_num_step_conf": 4.1015625,
      "calib/ece": 0.19094117647058823,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0010731707317073172,
      "calib/mean_conf": 0.005137254901960785,
      "calib/mu_c": 0.006,
      "calib/mu_w": 0.004926829268292683,
      "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": 0.00989557161373086,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 642.0,
      "completions/max_terminated_length": 642.0,
      "completions/mean_length": 186.28125,
      "completions/mean_terminated_length": 187.01177978515625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 71.0,
      "epoch": 0.05226666666666667,
      "grad_norm": 0.011714048683643341,
      "learning_rate": 4.194444444444445e-06,
      "loss": 0.0393,
      "num_tokens": 9800538.0,
      "reward": 0.9972037076950073,
      "reward_std": 0.013753028586506844,
      "rewards/accuracy_reward_step": 0.1953125,
      "rewards/final_brier_reward_step": 0.8030011653900146,
      "rewards/format_reward_step": 0.99609375,
      "step": 49
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4833144247531891,
      "aux_distill/mean_u": 0.3095816373875393,
      "aux_distill/n_active_tok": 267.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5081073896863371,
      "calib/avg_num_step_conf": 4.19140625,
      "calib/ece": 0.21882352941176472,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0006379585326953749,
      "calib/mean_conf": 0.004705882352941177,
      "calib/mu_c": 0.004210526315789474,
      "calib/mu_w": 0.0048484848484848485,
      "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": 0.007809732785584766,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 454.0,
      "completions/max_terminated_length": 454.0,
      "completions/mean_length": 190.09765625,
      "completions/mean_terminated_length": 190.84315490722656,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 54.0,
      "epoch": 0.05333333333333334,
      "grad_norm": 0.008671514689922333,
      "learning_rate": 4.166666666666667e-06,
      "loss": 0.0564,
      "num_tokens": 9954563.0,
      "reward": 0.9969898462295532,
      "reward_std": 0.012864282354712486,
      "rewards/accuracy_reward_step": 0.22265625,
      "rewards/final_brier_reward_step": 0.7752296924591064,
      "rewards/format_reward_step": 0.99609375,
      "step": 50
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.46149670518934727,
      "aux_distill/mean_u": 0.2551955135773979,
      "aux_distill/n_active_tok": 255.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5217850438047559,
      "calib/avg_num_step_conf": 4.015625,
      "calib/ece": 0.2618359375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0008479349186483118,
      "calib/mean_conf": 0.0037890625000000003,
      "calib/mu_c": 0.004411764705882354,
      "calib/mu_w": 0.003563829787234042,
      "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": 0.0064422341133409415,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 480.0,
      "completions/max_terminated_length": 480.0,
      "completions/mean_length": 180.6328125,
      "completions/mean_terminated_length": 181.3411865234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 73.0,
      "epoch": 0.0544,
      "grad_norm": 0.009305499494075775,
      "learning_rate": 4.138888888888889e-06,
      "loss": 0.0442,
      "num_tokens": 10110101.0,
      "reward": 1.0011439323425293,
      "reward_std": 0.0023848102428019047,
      "rewards/accuracy_reward_step": 0.265625,
      "rewards/final_brier_reward_step": 0.7366628646850586,
      "rewards/format_reward_step": 1.0,
      "step": 51
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48208334017544985,
      "aux_distill/mean_u": 0.31483845365810226,
      "aux_distill/n_active_tok": 242.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.48001508295625944,
      "calib/avg_num_step_conf": 3.80078125,
      "calib/ece": 0.1998828125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0006900452488687778,
      "calib/mean_conf": 0.0032421875000000003,
      "calib/mu_c": 0.0026923076923076926,
      "calib/mu_w": 0.0033823529411764705,
      "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": 0.005799038300860216,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 527.0,
      "completions/max_terminated_length": 527.0,
      "completions/mean_length": 179.19140625,
      "completions/mean_terminated_length": 179.89413452148438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 77.0,
      "epoch": 0.055466666666666664,
      "grad_norm": 0.008505421690642834,
      "learning_rate": 4.111111111111111e-06,
      "loss": 0.045,
      "num_tokens": 10263926.0,
      "reward": 1.0005247592926025,
      "reward_std": 0.0010291299549862742,
      "rewards/accuracy_reward_step": 0.203125,
      "rewards/final_brier_reward_step": 0.7979245781898499,
      "rewards/format_reward_step": 1.0,
      "step": 52
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4751516552641988,
      "aux_distill/mean_u": 0.3170315844694307,
      "aux_distill/n_active_tok": 251.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5117115737905695,
      "calib/avg_num_step_conf": 3.94921875,
      "calib/ece": 0.2751764705882353,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00017375995101041007,
      "calib/mean_conf": 0.003254901960784314,
      "calib/mu_c": 0.0033802816901408453,
      "calib/mu_w": 0.003206521739130435,
      "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": 0.005386414054167887,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 481.0,
      "completions/max_terminated_length": 481.0,
      "completions/mean_length": 184.03125,
      "completions/mean_terminated_length": 184.75294494628906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 66.0,
      "epoch": 0.05653333333333333,
      "grad_norm": 0.008253109641373158,
      "learning_rate": 4.083333333333334e-06,
      "loss": 0.0498,
      "num_tokens": 10416862.0,
      "reward": 0.9970114827156067,
      "reward_std": 0.012826445512473583,
      "rewards/accuracy_reward_step": 0.27734375,
      "rewards/final_brier_reward_step": 0.7205855250358582,
      "rewards/format_reward_step": 0.99609375,
      "step": 53
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.46900024451315403,
      "aux_distill/mean_u": 0.26208520407903335,
      "aux_distill/n_active_tok": 254.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5501342233185809,
      "calib/avg_num_step_conf": 3.99609375,
      "calib/ece": 0.29734375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.001398824639048103,
      "calib/mean_conf": 0.0034375000000000005,
      "calib/mu_c": 0.004415584415584416,
      "calib/mu_w": 0.003016759776536313,
      "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": 0.006547602137424051,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 527.0,
      "completions/max_terminated_length": 527.0,
      "completions/mean_length": 181.4609375,
      "completions/mean_terminated_length": 182.1725616455078,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 51.0,
      "epoch": 0.0576,
      "grad_norm": 0.0095495181158185,
      "learning_rate": 4.055555555555556e-06,
      "loss": 0.0294,
      "num_tokens": 10569548.0,
      "reward": 0.9973945021629333,
      "reward_std": 0.013665013015270233,
      "rewards/accuracy_reward_step": 0.30078125,
      "rewards/final_brier_reward_step": 0.6979140043258667,
      "rewards/format_reward_step": 0.99609375,
      "step": 54
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4792042141780257,
      "aux_distill/mean_u": 0.3025881007964957,
      "aux_distill/n_active_tok": 244.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5178941141674062,
      "calib/avg_num_step_conf": 3.8203125,
      "calib/ece": 0.1881640625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.000533372769397614,
      "calib/mean_conf": 0.0032421875000000003,
      "calib/mu_c": 0.003673469387755102,
      "calib/mu_w": 0.003140096618357488,
      "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": 0.005523028627016499,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 555.0,
      "completions/max_terminated_length": 555.0,
      "completions/mean_length": 182.6015625,
      "completions/mean_terminated_length": 183.31765747070312,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 54.0,
      "epoch": 0.058666666666666666,
      "grad_norm": 0.008562219329178333,
      "learning_rate": 4.027777777777779e-06,
      "loss": 0.058,
      "num_tokens": 10724118.0,
      "reward": 1.0006825923919678,
      "reward_std": 0.001537383534014225,
      "rewards/accuracy_reward_step": 0.19140625,
      "rewards/final_brier_reward_step": 0.8099589943885803,
      "rewards/format_reward_step": 1.0,
      "step": 55
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4713021432980895,
      "aux_distill/mean_u": 0.2991845504353514,
      "aux_distill/n_active_tok": 249.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.518120737045709,
      "calib/avg_num_step_conf": 3.9140625,
      "calib/ece": 0.180078125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0011666496996844146,
      "calib/mean_conf": 0.003515625,
      "calib/mu_c": 0.004468085106382979,
      "calib/mu_w": 0.0033014354066985643,
      "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": 0.006071172115775915,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 449.0,
      "completions/max_terminated_length": 449.0,
      "completions/mean_length": 183.203125,
      "completions/mean_terminated_length": 183.9215850830078,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 68.0,
      "epoch": 0.05973333333333333,
      "grad_norm": 0.005701439920812845,
      "learning_rate": 4.000000000000001e-06,
      "loss": 0.0394,
      "num_tokens": 10877858.0,
      "reward": 1.000795602798462,
      "reward_std": 0.0013140874216333032,
      "rewards/accuracy_reward_step": 0.18359375,
      "rewards/final_brier_reward_step": 0.8179976344108582,
      "rewards/format_reward_step": 1.0,
      "step": 56
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47348556853830814,
      "aux_distill/mean_u": 0.2610358596491855,
      "aux_distill/n_active_tok": 250.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5510937376609019,
      "calib/avg_num_step_conf": 3.921875,
      "calib/ece": 0.2584375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0014183052989023143,
      "calib/mean_conf": 0.0032812500000000003,
      "calib/mu_c": 0.004328358208955225,
      "calib/mu_w": 0.0029100529100529104,
      "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": 0.006073684914242754,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 411.0,
      "completions/max_terminated_length": 411.0,
      "completions/mean_length": 181.84375,
      "completions/mean_terminated_length": 182.55686950683594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 48.0,
      "epoch": 0.0608,
      "grad_norm": 0.00831560418009758,
      "learning_rate": 3.972222222222223e-06,
      "loss": 0.0481,
      "num_tokens": 11031202.0,
      "reward": 1.001109004020691,
      "reward_std": 0.002224062103778124,
      "rewards/accuracy_reward_step": 0.26171875,
      "rewards/final_brier_reward_step": 0.7404991984367371,
      "rewards/format_reward_step": 1.0,
      "step": 57
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.45535885356366634,
      "aux_distill/mean_u": 0.25883576058026386,
      "aux_distill/n_active_tok": 226.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4745469190495369,
      "calib/avg_num_step_conf": 3.53515625,
      "calib/ece": 0.2484765625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0006790173177607731,
      "calib/mean_conf": 0.0054296875000000005,
      "calib/mu_c": 0.004923076923076923,
      "calib/mu_w": 0.005602094240837696,
      "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": 0.007487681126513318,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 480.0,
      "completions/max_terminated_length": 480.0,
      "completions/mean_length": 162.53125,
      "completions/mean_terminated_length": 163.16864013671875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 48.0,
      "epoch": 0.06186666666666667,
      "grad_norm": 0.011086608283221722,
      "learning_rate": 3.944444444444445e-06,
      "loss": 0.0392,
      "num_tokens": 11179130.0,
      "reward": 0.9973009824752808,
      "reward_std": 0.013514685444533825,
      "rewards/accuracy_reward_step": 0.25390625,
      "rewards/final_brier_reward_step": 0.7446019649505615,
      "rewards/format_reward_step": 0.99609375,
      "step": 58
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4486118769273162,
      "aux_distill/mean_u": 0.24863800890353696,
      "aux_distill/n_active_tok": 227.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5085217825074826,
      "calib/avg_num_step_conf": 3.5625,
      "calib/ece": 0.2369921875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00038077818423678123,
      "calib/mean_conf": 0.0051953125,
      "calib/mu_c": 0.005483870967741936,
      "calib/mu_w": 0.005103092783505155,
      "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": 0.006956964354324646,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 512.0,
      "completions/max_terminated_length": 512.0,
      "completions/mean_length": 160.734375,
      "completions/mean_terminated_length": 161.36471557617188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 51.0,
      "epoch": 0.06293333333333333,
      "grad_norm": 0.008816853165626526,
      "learning_rate": 3.916666666666667e-06,
      "loss": 0.0604,
      "num_tokens": 11326526.0,
      "reward": 1.0012903213500977,
      "reward_std": 0.0024084849283099174,
      "rewards/accuracy_reward_step": 0.2421875,
      "rewards/final_brier_reward_step": 0.7603933811187744,
      "rewards/format_reward_step": 1.0,
      "step": 59
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47772796731442213,
      "aux_distill/mean_u": 0.25235217135883486,
      "aux_distill/n_active_tok": 194.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5245481294661622,
      "calib/avg_num_step_conf": 3.05078125,
      "calib/ece": 0.2324609375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0007524169819251789,
      "calib/mean_conf": 0.0058203125,
      "calib/mu_c": 0.00639344262295082,
      "calib/mu_w": 0.0056410256410256415,
      "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": 0.0068544392478410475,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 401.0,
      "completions/max_terminated_length": 401.0,
      "completions/mean_length": 143.34765625,
      "completions/mean_terminated_length": 143.90980529785156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 48.0,
      "epoch": 0.064,
      "grad_norm": 0.012299815192818642,
      "learning_rate": 3.88888888888889e-06,
      "loss": 0.039,
      "num_tokens": 11472079.0,
      "reward": 0.9975767135620117,
      "reward_std": 0.013632569462060928,
      "rewards/accuracy_reward_step": 0.23828125,
      "rewards/final_brier_reward_step": 0.7607784867286682,
      "rewards/format_reward_step": 0.99609375,
      "step": 60
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47281746193766594,
      "aux_distill/mean_u": 0.2551864848975026,
      "aux_distill/n_active_tok": 169.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.45449197120708745,
      "calib/avg_num_step_conf": 2.65625,
      "calib/ece": 0.3207421875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.001065891472868217,
      "calib/mean_conf": 0.0073828125000000005,
      "calib/mu_c": 0.006666666666666667,
      "calib/mu_w": 0.007732558139534884,
      "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": 0.007892461883965215,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 255.0,
      "completions/max_terminated_length": 255.0,
      "completions/mean_length": 123.0390625,
      "completions/mean_terminated_length": 123.52157592773438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.06506666666666666,
      "grad_norm": 0.012310824356973171,
      "learning_rate": 3.861111111111112e-06,
      "loss": 0.0563,
      "num_tokens": 11607641.0,
      "reward": 1.002129077911377,
      "reward_std": 0.003747039008885622,
      "rewards/accuracy_reward_step": 0.328125,
      "rewards/final_brier_reward_step": 0.6761331558227539,
      "rewards/format_reward_step": 1.0,
      "step": 61
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47183617390692234,
      "aux_distill/mean_u": 0.25694304938121587,
      "aux_distill/n_active_tok": 170.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.3787030438766161,
      "calib/avg_num_step_conf": 2.66015625,
      "calib/ece": 0.17484375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.003420543622111371,
      "calib/mean_conf": 0.00875,
      "calib/mu_c": 0.005957446808510639,
      "calib/mu_w": 0.00937799043062201,
      "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": 0.0082915619758885,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 294.0,
      "completions/max_terminated_length": 294.0,
      "completions/mean_length": 128.25,
      "completions/mean_terminated_length": 128.75294494628906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 45.0,
      "epoch": 0.06613333333333334,
      "grad_norm": 0.010938399471342564,
      "learning_rate": 3.833333333333334e-06,
      "loss": 0.0424,
      "num_tokens": 11747553.0,
      "reward": 1.0010210275650024,
      "reward_std": 0.002157938200980425,
      "rewards/accuracy_reward_step": 0.18359375,
      "rewards/final_brier_reward_step": 0.8184484243392944,
      "rewards/format_reward_step": 1.0,
      "step": 62
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.49205196741968393,
      "aux_distill/mean_u": 0.30069891272887367,
      "aux_distill/n_active_tok": 155.78125,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5284027282907462,
      "calib/avg_num_step_conf": 2.4375,
      "calib/ece": 0.174609375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0007227934439580589,
      "calib/mean_conf": 0.008984375,
      "calib/mu_c": 0.009574468085106385,
      "calib/mu_w": 0.008851674641148326,
      "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": 0.008181519165740246,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 274.0,
      "completions/max_terminated_length": 274.0,
      "completions/mean_length": 116.296875,
      "completions/mean_terminated_length": 116.75294494628906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 48.0,
      "epoch": 0.0672,
      "grad_norm": 0.011508263647556305,
      "learning_rate": 3.8055555555555556e-06,
      "loss": 0.0446,
      "num_tokens": 11885965.0,
      "reward": 1.0016839504241943,
      "reward_std": 0.002663816325366497,
      "rewards/accuracy_reward_step": 0.18359375,
      "rewards/final_brier_reward_step": 0.8197742104530334,
      "rewards/format_reward_step": 1.0,
      "step": 63
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4663435900583863,
      "aux_distill/mean_u": 0.2517591965568636,
      "aux_distill/n_active_tok": 150.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.44442207007652035,
      "calib/avg_num_step_conf": 2.35546875,
      "calib/ece": 0.2445703125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0017905759162303667,
      "calib/mean_conf": 0.009335937499999999,
      "calib/mu_c": 0.008,
      "calib/mu_w": 0.009790575916230367,
      "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": 0.00809781736001089,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 280.0,
      "completions/max_terminated_length": 280.0,
      "completions/mean_length": 115.34375,
      "completions/mean_terminated_length": 115.79608154296875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.06826666666666667,
      "grad_norm": 0.012494472786784172,
      "learning_rate": 3.777777777777778e-06,
      "loss": 0.044,
      "num_tokens": 12019269.0,
      "reward": 1.0019547939300537,
      "reward_std": 0.0032990314066410065,
      "rewards/accuracy_reward_step": 0.25390625,
      "rewards/final_brier_reward_step": 0.7500035762786865,
      "rewards/format_reward_step": 1.0,
      "step": 64
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48996405489742756,
      "aux_distill/mean_u": 0.25722923283434773,
      "aux_distill/n_active_tok": 134.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.49447898799313894,
      "calib/avg_num_step_conf": 2.1015625,
      "calib/ece": 0.16171875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00018867924528301674,
      "calib/mean_conf": 0.010156249999999999,
      "calib/mu_c": 0.010000000000000002,
      "calib/mu_w": 0.010188679245283019,
      "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": 0.008383799015810196,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 244.0,
      "completions/max_terminated_length": 244.0,
      "completions/mean_length": 110.921875,
      "completions/mean_terminated_length": 111.35687255859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.06933333333333333,
      "grad_norm": 0.01302141323685646,
      "learning_rate": 3.7500000000000005e-06,
      "loss": 0.0479,
      "num_tokens": 12152689.0,
      "reward": 1.0016319751739502,
      "reward_std": 0.0025428631342947483,
      "rewards/accuracy_reward_step": 0.171875,
      "rewards/final_brier_reward_step": 0.8313890695571899,
      "rewards/format_reward_step": 1.0,
      "step": 65
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4909752355888486,
      "aux_distill/mean_u": 0.26345752851416304,
      "aux_distill/n_active_tok": 152.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5645089285714286,
      "calib/avg_num_step_conf": 2.390625,
      "calib/ece": 0.2095703125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.002192857142857144,
      "calib/mean_conf": 0.009179687499999999,
      "calib/mu_c": 0.010892857142857143,
      "calib/mu_w": 0.0087,
      "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": 0.00753319071856964,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 321.0,
      "completions/max_terminated_length": 321.0,
      "completions/mean_length": 121.94921875,
      "completions/mean_terminated_length": 122.42745971679688,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.0704,
      "grad_norm": 0.013737301342189312,
      "learning_rate": 3.7222222222222225e-06,
      "loss": 0.045,
      "num_tokens": 12290260.0,
      "reward": 1.0023123025894165,
      "reward_std": 0.0038235776592046022,
      "rewards/accuracy_reward_step": 0.21875,
      "rewards/final_brier_reward_step": 0.785874605178833,
      "rewards/format_reward_step": 1.0,
      "step": 66
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4867994925007224,
      "aux_distill/mean_u": 0.2638074353122962,
      "aux_distill/n_active_tok": 157.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5015463917525774,
      "calib/avg_num_step_conf": 2.4609375,
      "calib/ece": 0.22720472440944883,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0009793814432989666,
      "calib/mean_conf": 0.010748031496062993,
      "calib/mu_c": 0.010000000000000002,
      "calib/mu_w": 0.010979381443298969,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0008661417322834645,
      "calib/std_conf": 0.014974765390055263,
      "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": 1753.0,
      "completions/max_terminated_length": 1753.0,
      "completions/mean_length": 136.06640625,
      "completions/mean_terminated_length": 136.06640625,
      "completions/min_length": 49.0,
      "completions/min_terminated_length": 49.0,
      "epoch": 0.07146666666666666,
      "grad_norm": 0.012562266550958157,
      "learning_rate": 3.694444444444445e-06,
      "loss": 0.0889,
      "num_tokens": 12430101.0,
      "reward": 0.9963157773017883,
      "reward_std": 0.019671740010380745,
      "rewards/accuracy_reward_step": 0.23828125,
      "rewards/final_brier_reward_step": 0.7621628642082214,
      "rewards/format_reward_step": 0.9921875,
      "step": 67
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.469187312759459,
      "aux_distill/mean_u": 0.24902920117743751,
      "aux_distill/n_active_tok": 162.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5078149920255183,
      "calib/avg_num_step_conf": 2.54296875,
      "calib/ece": 0.24796875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0006188197767145118,
      "calib/mean_conf": 0.009843750000000002,
      "calib/mu_c": 0.010303030303030302,
      "calib/mu_w": 0.00968421052631579,
      "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": 0.007179003129787589,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 411.0,
      "completions/max_terminated_length": 411.0,
      "completions/mean_length": 129.4921875,
      "completions/mean_terminated_length": 130.00001525878906,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 54.0,
      "epoch": 0.07253333333333334,
      "grad_norm": 0.01084089931100607,
      "learning_rate": 3.6666666666666666e-06,
      "loss": 0.0456,
      "num_tokens": 12567339.0,
      "reward": 1.00258207321167,
      "reward_std": 0.003189860377460718,
      "rewards/accuracy_reward_step": 0.2578125,
      "rewards/final_brier_reward_step": 0.7473515272140503,
      "rewards/format_reward_step": 1.0,
      "step": 68
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4741177558898926,
      "aux_distill/mean_u": 0.2582151062249327,
      "aux_distill/n_active_tok": 153.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.468495145631068,
      "calib/avg_num_step_conf": 2.421875,
      "calib/ece": 0.1847265625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0007281553398058253,
      "calib/mean_conf": 0.0105859375,
      "calib/mu_c": 0.01,
      "calib/mu_w": 0.010728155339805825,
      "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": 0.007074411795060685,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 340.0,
      "completions/max_terminated_length": 340.0,
      "completions/mean_length": 129.60546875,
      "completions/mean_terminated_length": 130.11373901367188,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 47.0,
      "epoch": 0.0736,
      "grad_norm": 0.013099942356348038,
      "learning_rate": 3.638888888888889e-06,
      "loss": 0.0516,
      "num_tokens": 12705014.0,
      "reward": 1.0018720626831055,
      "reward_std": 0.002898347331210971,
      "rewards/accuracy_reward_step": 0.1953125,
      "rewards/final_brier_reward_step": 0.8084316253662109,
      "rewards/format_reward_step": 1.0,
      "step": 69
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.493006375618279,
      "aux_distill/mean_u": 0.262141115781891,
      "aux_distill/n_active_tok": 160.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.554221931627125,
      "calib/avg_num_step_conf": 2.5,
      "calib/ece": 0.198,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.001150756585092472,
      "calib/mean_conf": 0.009843137254901962,
      "calib/mu_c": 0.010754716981132076,
      "calib/mu_w": 0.009603960396039604,
      "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": 0.007354797151385198,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 372.0,
      "completions/max_terminated_length": 372.0,
      "completions/mean_length": 133.13671875,
      "completions/mean_terminated_length": 133.65882873535156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 50.0,
      "epoch": 0.07466666666666667,
      "grad_norm": 0.014240157790482044,
      "learning_rate": 3.6111111111111115e-06,
      "loss": 0.0264,
      "num_tokens": 12846089.0,
      "reward": 0.9982450008392334,
      "reward_std": 0.013962855562567711,
      "rewards/accuracy_reward_step": 0.20703125,
      "rewards/final_brier_reward_step": 0.7933652400970459,
      "rewards/format_reward_step": 0.99609375,
      "step": 70
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47487139515578747,
      "aux_distill/mean_u": 0.27355696705036336,
      "aux_distill/n_active_tok": 168.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.48920236851271337,
      "calib/avg_num_step_conf": 2.63671875,
      "calib/ece": 0.21640625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00020202020202019985,
      "calib/mean_conf": 0.010156249999999999,
      "calib/mu_c": 0.010000000000000002,
      "calib/mu_w": 0.010202020202020202,
      "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": 0.007601847534481339,
      "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": 377.0,
      "completions/max_terminated_length": 377.0,
      "completions/mean_length": 144.0546875,
      "completions/mean_terminated_length": 144.61961364746094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 52.0,
      "epoch": 0.07573333333333333,
      "grad_norm": 0.010125796310603619,
      "learning_rate": 3.5833333333333335e-06,
      "loss": 0.0251,
      "num_tokens": 12987375.0,
      "reward": 0.9982790946960449,
      "reward_std": 0.013809540309011936,
      "rewards/accuracy_reward_step": 0.2265625,
      "rewards/final_brier_reward_step": 0.7739019393920898,
      "rewards/format_reward_step": 0.99609375,
      "step": 71
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48494631331413984,
      "aux_distill/mean_u": 0.26469371248366147,
      "aux_distill/n_active_tok": 172.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5170990566037735,
      "calib/avg_num_step_conf": 2.71875,
      "calib/ece": 0.16125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0006174957118353329,
      "calib/mean_conf": 0.010625,
      "calib/mu_c": 0.011136363636363637,
      "calib/mu_w": 0.010518867924528304,
      "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": 0.006643841132959155,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 331.0,
      "completions/max_terminated_length": 331.0,
      "completions/mean_length": 140.33203125,
      "completions/mean_terminated_length": 140.88235473632812,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 55.0,
      "epoch": 0.0768,
      "grad_norm": 0.011124787852168083,
      "learning_rate": 3.555555555555556e-06,
      "loss": 0.0294,
      "num_tokens": 13127708.0,
      "reward": 1.001835584640503,
      "reward_std": 0.002578023122623563,
      "rewards/accuracy_reward_step": 0.171875,
      "rewards/final_brier_reward_step": 0.8317961096763611,
      "rewards/format_reward_step": 1.0,
      "step": 72
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4593647988513112,
      "aux_distill/mean_u": 0.2135954470103809,
      "aux_distill/n_active_tok": 163.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.44977912113461993,
      "calib/avg_num_step_conf": 2.578125,
      "calib/ece": 0.2582421875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.001169495466170659,
      "calib/mean_conf": 0.0112890625,
      "calib/mu_c": 0.010434782608695651,
      "calib/mu_w": 0.01160427807486631,
      "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": 0.006019048335999117,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 422.0,
      "completions/max_terminated_length": 422.0,
      "completions/mean_length": 141.09375,
      "completions/mean_terminated_length": 141.64706420898438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 46.0,
      "epoch": 0.07786666666666667,
      "grad_norm": 0.012260450050234795,
      "learning_rate": 3.5277777777777784e-06,
      "loss": 0.0498,
      "num_tokens": 13270860.0,
      "reward": 1.0027306079864502,
      "reward_std": 0.0034748294856399298,
      "rewards/accuracy_reward_step": 0.26953125,
      "rewards/final_brier_reward_step": 0.7359300851821899,
      "rewards/format_reward_step": 1.0,
      "step": 73
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47802817448973656,
      "aux_distill/mean_u": 0.22195564799117634,
      "aux_distill/n_active_tok": 156.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5704526442977972,
      "calib/avg_num_step_conf": 2.4453125,
      "calib/ece": 0.1956640625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0018449670043684356,
      "calib/mean_conf": 0.0113671875,
      "calib/mu_c": 0.012830188679245284,
      "calib/mu_w": 0.010985221674876849,
      "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": 0.005870172343282925,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 364.0,
      "completions/max_terminated_length": 364.0,
      "completions/mean_length": 131.05859375,
      "completions/mean_terminated_length": 131.5725555419922,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 51.0,
      "epoch": 0.07893333333333333,
      "grad_norm": 0.011557850986719131,
      "learning_rate": 3.5e-06,
      "loss": 0.0515,
      "num_tokens": 13408339.0,
      "reward": 1.0025744438171387,
      "reward_std": 0.0030165817588567734,
      "rewards/accuracy_reward_step": 0.20703125,
      "rewards/final_brier_reward_step": 0.7981175184249878,
      "rewards/format_reward_step": 1.0,
      "step": 74
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.45104537066072226,
      "aux_distill/mean_u": 0.23226746896270406,
      "aux_distill/n_active_tok": 157.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4995579133510168,
      "calib/avg_num_step_conf": 2.46875,
      "calib/ece": 0.3296875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -6.257226416377278e-05,
      "calib/mean_conf": 0.01015625,
      "calib/mu_c": 0.010114942528735633,
      "calib/mu_w": 0.010177514792899406,
      "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": 0.005300998579277304,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 360.0,
      "completions/max_terminated_length": 360.0,
      "completions/mean_length": 134.49609375,
      "completions/mean_terminated_length": 135.02354431152344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 52.0,
      "epoch": 0.08,
      "grad_norm": 0.012841647490859032,
      "learning_rate": 3.4722222222222224e-06,
      "loss": 0.0255,
      "num_tokens": 13547522.0,
      "reward": 1.0033719539642334,
      "reward_std": 0.003874801332131028,
      "rewards/accuracy_reward_step": 0.33984375,
      "rewards/final_brier_reward_step": 0.6669000387191772,
      "rewards/format_reward_step": 1.0,
      "step": 75
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.48809313867241144,
      "aux_distill/mean_u": 0.2720469078054451,
      "aux_distill/n_active_tok": 149.78125,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5330114529530654,
      "calib/avg_num_step_conf": 2.35546875,
      "calib/ece": 0.2738671875,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0006879257429448296,
      "calib/mean_conf": 0.011289062500000002,
      "calib/mu_c": 0.011780821917808219,
      "calib/mu_w": 0.011092896174863389,
      "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": 0.005029059839681146,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 319.0,
      "completions/max_terminated_length": 319.0,
      "completions/mean_length": 126.2109375,
      "completions/mean_terminated_length": 126.70588684082031,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 55.0,
      "epoch": 0.08106666666666666,
      "grad_norm": 0.012701224535703659,
      "learning_rate": 3.444444444444445e-06,
      "loss": 0.043,
      "num_tokens": 13682888.0,
      "reward": 1.003282904624939,
      "reward_std": 0.003568597137928009,
      "rewards/accuracy_reward_step": 0.28515625,
      "rewards/final_brier_reward_step": 0.721409797668457,
      "rewards/format_reward_step": 1.0,
      "step": 76
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4653266780078411,
      "aux_distill/mean_u": 0.23650410068623975,
      "aux_distill/n_active_tok": 163.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.46630100556802123,
      "calib/avg_num_step_conf": 2.5546875,
      "calib/ece": 0.23669291338582676,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0007246738136790481,
      "calib/mean_conf": 0.011338582677165353,
      "calib/mu_c": 0.010793650793650796,
      "calib/mu_w": 0.011518324607329844,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.004837876270647444,
      "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": 966.0,
      "completions/max_terminated_length": 966.0,
      "completions/mean_length": 139.02734375,
      "completions/mean_terminated_length": 139.02734375,
      "completions/min_length": 57.0,
      "completions/min_terminated_length": 57.0,
      "epoch": 0.08213333333333334,
      "grad_norm": 0.01345721073448658,
      "learning_rate": 3.416666666666667e-06,
      "loss": 0.0603,
      "num_tokens": 13823143.0,
      "reward": 0.9967213869094849,
      "reward_std": 0.01972099393606186,
      "rewards/accuracy_reward_step": 0.25,
      "rewards/final_brier_reward_step": 0.7512555122375488,
      "rewards/format_reward_step": 0.9921875,
      "step": 77
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4758943486958742,
      "aux_distill/mean_u": 0.24784744747853024,
      "aux_distill/n_active_tok": 165.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5372364122364123,
      "calib/avg_num_step_conf": 2.59765625,
      "calib/ece": 0.277890625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0008226908226908226,
      "calib/mean_conf": 0.011171875000000001,
      "calib/mu_c": 0.011756756756756755,
      "calib/mu_w": 0.010934065934065932,
      "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": 0.0045278812908881565,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 372.0,
      "completions/max_terminated_length": 372.0,
      "completions/mean_length": 146.38671875,
      "completions/mean_terminated_length": 146.96080017089844,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 61.0,
      "epoch": 0.0832,
      "grad_norm": 0.013810593634843826,
      "learning_rate": 3.3888888888888893e-06,
      "loss": 0.0437,
      "num_tokens": 13968642.0,
      "reward": 1.0033257007598877,
      "reward_std": 0.004381269216537476,
      "rewards/accuracy_reward_step": 0.2890625,
      "rewards/final_brier_reward_step": 0.7175890207290649,
      "rewards/format_reward_step": 1.0,
      "step": 78
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.447107189334929,
      "aux_distill/mean_u": 0.21981514835849306,
      "aux_distill/n_active_tok": 172.75,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5005133064834557,
      "calib/avg_num_step_conf": 2.70703125,
      "calib/ece": 0.2505078125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -2.29013661849492e-05,
      "calib/mean_conf": 0.0112109375,
      "calib/mu_c": 0.011194029850746268,
      "calib/mu_w": 0.011216931216931217,
      "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": 0.00464548225387782,
      "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": 376.0,
      "completions/max_terminated_length": 376.0,
      "completions/mean_length": 149.1171875,
      "completions/mean_terminated_length": 149.70196533203125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 61.0,
      "epoch": 0.08426666666666667,
      "grad_norm": 0.01121280062943697,
      "learning_rate": 3.3611111111111117e-06,
      "loss": 0.0497,
      "num_tokens": 14113192.0,
      "reward": 0.9989499449729919,
      "reward_std": 0.013474896550178528,
      "rewards/accuracy_reward_step": 0.26171875,
      "rewards/final_brier_reward_step": 0.7400875091552734,
      "rewards/format_reward_step": 0.99609375,
      "step": 79
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4624016094021499,
      "aux_distill/mean_u": 0.25210709773324724,
      "aux_distill/n_active_tok": 186.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4720052083333333,
      "calib/avg_num_step_conf": 2.91015625,
      "calib/ece": 0.363359375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00045833333333333316,
      "calib/mean_conf": 0.011640625,
      "calib/mu_c": 0.011354166666666667,
      "calib/mu_w": 0.0118125,
      "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": 0.00512001949306592,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 331.0,
      "completions/max_terminated_length": 331.0,
      "completions/mean_length": 157.421875,
      "completions/mean_terminated_length": 158.0392303466797,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 52.0,
      "epoch": 0.08533333333333333,
      "grad_norm": 0.011426246725022793,
      "learning_rate": 3.3333333333333333e-06,
      "loss": 0.0547,
      "num_tokens": 14255652.0,
      "reward": 1.0041768550872803,
      "reward_std": 0.004204465076327324,
      "rewards/accuracy_reward_step": 0.375,
      "rewards/final_brier_reward_step": 0.6333538889884949,
      "rewards/format_reward_step": 1.0,
      "step": 80
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.47395670786499977,
      "aux_distill/mean_u": 0.2623399104299867,
      "aux_distill/n_active_tok": 191.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5220156555772995,
      "calib/avg_num_step_conf": 2.9921875,
      "calib/ece": 0.2749803921568627,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0004899894625922004,
      "calib/mean_conf": 0.011294117647058823,
      "calib/mu_c": 0.011643835616438355,
      "calib/mu_w": 0.011153846153846155,
      "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": 0.004460256379015822,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 421.0,
      "completions/max_terminated_length": 421.0,
      "completions/mean_length": 170.6015625,
      "completions/mean_terminated_length": 171.27059936523438,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 65.0,
      "epoch": 0.0864,
      "grad_norm": 0.0148467430844903,
      "learning_rate": 3.3055555555555558e-06,
      "loss": 0.0472,
      "num_tokens": 14405574.0,
      "reward": 0.9993405938148499,
      "reward_std": 0.014343726448714733,
      "rewards/accuracy_reward_step": 0.28515625,
      "rewards/final_brier_reward_step": 0.7174311876296997,
      "rewards/format_reward_step": 0.99609375,
      "step": 81
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.44830169156193733,
      "aux_distill/mean_u": 0.20830730432778605,
      "aux_distill/n_active_tok": 182.84375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5069738480697386,
      "calib/avg_num_step_conf": 2.859375,
      "calib/ece": 0.418125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00012951432129514572,
      "calib/mean_conf": 0.011562500000000002,
      "calib/mu_c": 0.011636363636363639,
      "calib/mu_w": 0.011506849315068493,
      "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": 0.004582149468317244,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 544.0,
      "completions/max_terminated_length": 544.0,
      "completions/mean_length": 175.63671875,
      "completions/mean_terminated_length": 176.32550048828125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 69.0,
      "epoch": 0.08746666666666666,
      "grad_norm": 0.011641280725598335,
      "learning_rate": 3.277777777777778e-06,
      "loss": 0.0327,
      "num_tokens": 14556089.0,
      "reward": 1.0010164976119995,
      "reward_std": 0.01543247140944004,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.5762519240379333,
      "rewards/format_reward_step": 0.99609375,
      "step": 82
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4452169854193926,
      "aux_distill/mean_u": 0.23544867615008755,
      "aux_distill/n_active_tok": 194.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.547075933864054,
      "calib/avg_num_step_conf": 3.07421875,
      "calib/ece": 0.26623529411764707,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0008603796693202673,
      "calib/mean_conf": 0.01219607843137255,
      "calib/mu_c": 0.012816901408450702,
      "calib/mu_w": 0.011956521739130435,
      "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": 0.004756912855562644,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 804.0,
      "completions/max_terminated_length": 804.0,
      "completions/mean_length": 187.96875,
      "completions/mean_terminated_length": 188.7058868408203,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 48.0,
      "epoch": 0.08853333333333334,
      "grad_norm": 0.009983416646718979,
      "learning_rate": 3.2500000000000002e-06,
      "loss": 0.0472,
      "num_tokens": 14711473.0,
      "reward": 0.9995629787445068,
      "reward_std": 0.014329729601740837,
      "rewards/accuracy_reward_step": 0.27734375,
      "rewards/final_brier_reward_step": 0.7256886959075928,
      "rewards/format_reward_step": 0.99609375,
      "step": 83
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4570365957915783,
      "aux_distill/mean_u": 0.22510164656623236,
      "aux_distill/n_active_tok": 177.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5138473053892214,
      "calib/avg_num_step_conf": 2.76953125,
      "calib/ece": 0.3325490196078431,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0002721829069134464,
      "calib/mean_conf": 0.012549019607843138,
      "calib/mu_c": 0.012727272727272728,
      "calib/mu_w": 0.012455089820359281,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.004868118473312227,
      "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": 2196.0,
      "completions/max_terminated_length": 2196.0,
      "completions/mean_length": 175.78125,
      "completions/mean_terminated_length": 175.78125,
      "completions/min_length": 55.0,
      "completions/min_terminated_length": 55.0,
      "epoch": 0.0896,
      "grad_norm": 0.0139435064047575,
      "learning_rate": 3.2222222222222227e-06,
      "loss": 0.098,
      "num_tokens": 14862393.0,
      "reward": 1.0003783702850342,
      "reward_std": 0.015527335926890373,
      "rewards/accuracy_reward_step": 0.34375,
      "rewards/final_brier_reward_step": 0.6609132885932922,
      "rewards/format_reward_step": 0.99609375,
      "step": 84
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.42052019108086824,
      "aux_distill/mean_u": 0.19761002619786422,
      "aux_distill/n_active_tok": 186.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.512276367991616,
      "calib/avg_num_step_conf": 2.921875,
      "calib/ece": 0.2724609375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 6.213039898195737e-05,
      "calib/mean_conf": 0.0126953125,
      "calib/mu_c": 0.01273972602739726,
      "calib/mu_w": 0.012677595628415302,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.005318097923820485,
      "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": 506.0,
      "completions/max_terminated_length": 506.0,
      "completions/mean_length": 189.62890625,
      "completions/mean_terminated_length": 190.37255859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 65.0,
      "epoch": 0.09066666666666667,
      "grad_norm": 0.012380489148199558,
      "learning_rate": 3.1944444444444443e-06,
      "loss": 0.0634,
      "num_tokens": 15018762.0,
      "reward": 0.9996320009231567,
      "reward_std": 0.015054484829306602,
      "rewards/accuracy_reward_step": 0.28515625,
      "rewards/final_brier_reward_step": 0.718014121055603,
      "rewards/format_reward_step": 0.99609375,
      "step": 85
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4251967119053006,
      "aux_distill/mean_u": 0.19627234303921945,
      "aux_distill/n_active_tok": 198.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5002316602316603,
      "calib/avg_num_step_conf": 3.10546875,
      "calib/ece": 0.26207843137254905,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00020077220077219932,
      "calib/mean_conf": 0.012431372549019607,
      "calib/mu_c": 0.012285714285714287,
      "calib/mu_w": 0.012486486486486486,
      "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": 0.005199000288369114,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2301.0,
      "completions/max_terminated_length": 2301.0,
      "completions/mean_length": 206.75390625,
      "completions/mean_terminated_length": 206.75390625,
      "completions/min_length": 47.0,
      "completions/min_terminated_length": 47.0,
      "epoch": 0.09173333333333333,
      "grad_norm": 0.010361351072788239,
      "learning_rate": 3.1666666666666667e-06,
      "loss": 0.0892,
      "num_tokens": 15177203.0,
      "reward": 0.9993627071380615,
      "reward_std": 0.014718771912157536,
      "rewards/accuracy_reward_step": 0.2734375,
      "rewards/final_brier_reward_step": 0.729194164276123,
      "rewards/format_reward_step": 0.99609375,
      "step": 86
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4048550510779023,
      "aux_distill/mean_u": 0.1747493499899473,
      "aux_distill/n_active_tok": 162.75,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.48597391084093206,
      "calib/avg_num_step_conf": 2.54296875,
      "calib/ece": 0.4300790513833992,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -3.482776089159048e-05,
      "calib/mean_conf": 0.012608695652173915,
      "calib/mu_c": 0.012589285714285714,
      "calib/mu_w": 0.012624113475177305,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.0056506534896365365,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2595.0,
      "completions/max_terminated_length": 2595.0,
      "completions/mean_length": 206.109375,
      "completions/mean_terminated_length": 206.109375,
      "completions/min_length": 78.0,
      "completions/min_terminated_length": 78.0,
      "epoch": 0.0928,
      "grad_norm": 0.01382330060005188,
      "learning_rate": 3.138888888888889e-06,
      "loss": 0.1513,
      "num_tokens": 15335463.0,
      "reward": 0.9897886514663696,
      "reward_std": 0.04076341167092323,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/final_brier_reward_step": 0.5577023029327393,
      "rewards/format_reward_step": 0.984375,
      "step": 87
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.40625389851629734,
      "aux_distill/mean_u": 0.17802966193899944,
      "aux_distill/n_active_tok": 179.75,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5151803324880248,
      "calib/avg_num_step_conf": 2.80859375,
      "calib/ece": 0.32,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0003684136376444075,
      "calib/mean_conf": 0.012015810276679842,
      "calib/mu_c": 0.012261904761904762,
      "calib/mu_w": 0.011893491124260355,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.004734841324391676,
      "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": 1933.0,
      "completions/max_terminated_length": 1933.0,
      "completions/mean_length": 229.66015625,
      "completions/mean_terminated_length": 230.560791015625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 68.0,
      "epoch": 0.09386666666666667,
      "grad_norm": 0.01094108261168003,
      "learning_rate": 3.1111111111111116e-06,
      "loss": 0.093,
      "num_tokens": 15504104.0,
      "reward": 0.9922221899032593,
      "reward_std": 0.03676649183034897,
      "rewards/accuracy_reward_step": 0.328125,
      "rewards/final_brier_reward_step": 0.6680383086204529,
      "rewards/format_reward_step": 0.98828125,
      "step": 88
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.42246151296421885,
      "aux_distill/mean_u": 0.21081985870031994,
      "aux_distill/n_active_tok": 167.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5588314124899492,
      "calib/avg_num_step_conf": 2.61328125,
      "calib/ece": 0.3439607843137255,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0012972393460198364,
      "calib/mean_conf": 0.012901960784313726,
      "calib/mu_c": 0.01373626373626374,
      "calib/mu_w": 0.012439024390243903,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.004951747174169285,
      "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": 1833.0,
      "completions/max_terminated_length": 1833.0,
      "completions/mean_length": 217.08984375,
      "completions/mean_terminated_length": 217.08984375,
      "completions/min_length": 68.0,
      "completions/min_terminated_length": 68.0,
      "epoch": 0.09493333333333333,
      "grad_norm": 0.012723850086331367,
      "learning_rate": 3.0833333333333336e-06,
      "loss": 0.0679,
      "num_tokens": 15668567.0,
      "reward": 0.9969753623008728,
      "reward_std": 0.02681366354227066,
      "rewards/accuracy_reward_step": 0.35546875,
      "rewards/final_brier_reward_step": 0.6462945342063904,
      "rewards/format_reward_step": 0.9921875,
      "step": 89
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.41852833703160286,
      "aux_distill/mean_u": 0.2219748970620147,
      "aux_distill/n_active_tok": 173.5625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.4908326967150497,
      "calib/avg_num_step_conf": 2.73046875,
      "calib/ece": 0.385390625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00017570664629488163,
      "calib/mean_conf": 0.013046875,
      "calib/mu_c": 0.012941176470588235,
      "calib/mu_w": 0.013116883116883117,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.004850675492586058,
      "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": 546.0,
      "completions/max_terminated_length": 546.0,
      "completions/mean_length": 208.47265625,
      "completions/mean_terminated_length": 209.2902069091797,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 78.0,
      "epoch": 0.096,
      "grad_norm": 0.011926371604204178,
      "learning_rate": 3.055555555555556e-06,
      "loss": 0.0544,
      "num_tokens": 15825256.0,
      "reward": 0.9991612434387207,
      "reward_std": 0.021257756277918816,
      "rewards/accuracy_reward_step": 0.3984375,
      "rewards/final_brier_reward_step": 0.607697606086731,
      "rewards/format_reward_step": 0.9921875,
      "step": 90
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38081121258437634,
      "aux_distill/mean_u": 0.160952301445108,
      "aux_distill/n_active_tok": 175.0625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5157251019219569,
      "calib/avg_num_step_conf": 2.734375,
      "calib/ece": 0.3838582677165355,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0001358959425354321,
      "calib/mean_conf": 0.013779527559055118,
      "calib/mu_c": 0.013861386138613863,
      "calib/mu_w": 0.013725490196078431,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.005531461036120781,
      "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": 2249.0,
      "completions/max_terminated_length": 2249.0,
      "completions/mean_length": 240.15625,
      "completions/mean_terminated_length": 240.15625,
      "completions/min_length": 88.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.09706666666666666,
      "grad_norm": 0.012279514223337173,
      "learning_rate": 3.0277777777777776e-06,
      "loss": 0.1304,
      "num_tokens": 15994448.0,
      "reward": 0.9975468516349792,
      "reward_std": 0.02685142122209072,
      "rewards/accuracy_reward_step": 0.39453125,
      "rewards/final_brier_reward_step": 0.6083749532699585,
      "rewards/format_reward_step": 0.9921875,
      "step": 91
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4037033971399069,
      "aux_distill/mean_u": 0.150256089948064,
      "aux_distill/n_active_tok": 164.5,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5810709604942895,
      "calib/avg_num_step_conf": 2.57421875,
      "calib/ece": 0.4109765625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0016981838607002427,
      "calib/mean_conf": 0.0148046875,
      "calib/mu_c": 0.015779816513761466,
      "calib/mu_w": 0.014081632653061223,
      "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": 0.005586756932903359,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 849.0,
      "completions/max_terminated_length": 849.0,
      "completions/mean_length": 203.796875,
      "completions/mean_terminated_length": 204.59608459472656,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.09813333333333334,
      "grad_norm": 0.014511768706142902,
      "learning_rate": 3e-06,
      "loss": 0.0525,
      "num_tokens": 16153340.0,
      "reward": 1.0065934658050537,
      "reward_std": 0.006135149858891964,
      "rewards/accuracy_reward_step": 0.42578125,
      "rewards/final_brier_reward_step": 0.5874058604240417,
      "rewards/format_reward_step": 1.0,
      "step": 92
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3932137689553201,
      "aux_distill/mean_u": 0.17165799268354254,
      "aux_distill/n_active_tok": 188.8125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.49097179955366205,
      "calib/avg_num_step_conf": 2.953125,
      "calib/ece": 0.3535714285714286,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00015824710894704955,
      "calib/mean_conf": 0.015476190476190473,
      "calib/mu_c": 0.015376344086021504,
      "calib/mu_w": 0.015534591194968554,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.005788212276734865,
      "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": 2072.0,
      "completions/max_terminated_length": 2072.0,
      "completions/mean_length": 254.62109375,
      "completions/mean_terminated_length": 255.61962890625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 87.0,
      "epoch": 0.0992,
      "grad_norm": 0.010879608802497387,
      "learning_rate": 2.9722222222222225e-06,
      "loss": 0.0918,
      "num_tokens": 16324299.0,
      "reward": 0.989826500415802,
      "reward_std": 0.048621803522109985,
      "rewards/accuracy_reward_step": 0.36328125,
      "rewards/final_brier_reward_step": 0.6319968700408936,
      "rewards/format_reward_step": 0.984375,
      "step": 93
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33277420699596405,
      "aux_distill/mean_u": 0.15313231940754068,
      "aux_distill/n_active_tok": 200.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6165608965703484,
      "calib/avg_num_step_conf": 3.125,
      "calib/ece": 0.34707509881422927,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0024999999999999988,
      "calib/mean_conf": 0.016561264822134384,
      "calib/mu_c": 0.018152173913043478,
      "calib/mu_w": 0.01565217391304348,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.005933858230446628,
      "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": 2009.0,
      "completions/max_terminated_length": 2009.0,
      "completions/mean_length": 247.44140625,
      "completions/mean_terminated_length": 247.44140625,
      "completions/min_length": 98.0,
      "completions/min_terminated_length": 98.0,
      "epoch": 0.10026666666666667,
      "grad_norm": 0.011224609799683094,
      "learning_rate": 2.944444444444445e-06,
      "loss": 0.1157,
      "num_tokens": 16496324.0,
      "reward": 0.994651734828949,
      "reward_std": 0.03712940961122513,
      "rewards/accuracy_reward_step": 0.359375,
      "rewards/final_brier_reward_step": 0.641647219657898,
      "rewards/format_reward_step": 0.98828125,
      "step": 94
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3890888011083007,
      "aux_distill/mean_u": 0.201429254871052,
      "aux_distill/n_active_tok": 186.4375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5229846153846154,
      "calib/avg_num_step_conf": 2.89453125,
      "calib/ece": 0.4926274509803921,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0005815384615384701,
      "calib/mean_conf": 0.017176470588235297,
      "calib/mu_c": 0.017461538461538466,
      "calib/mu_w": 0.016879999999999996,
      "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": 0.005728175572238498,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2106.0,
      "completions/max_terminated_length": 2106.0,
      "completions/mean_length": 236.3125,
      "completions/mean_terminated_length": 236.3125,
      "completions/min_length": 86.0,
      "completions/min_terminated_length": 86.0,
      "epoch": 0.10133333333333333,
      "grad_norm": 0.011299879290163517,
      "learning_rate": 2.916666666666667e-06,
      "loss": 0.1178,
      "num_tokens": 16662948.0,
      "reward": 1.0047976970672607,
      "reward_std": 0.016820628196001053,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.5056890249252319,
      "rewards/format_reward_step": 0.99609375,
      "step": 95
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37238421756774187,
      "aux_distill/mean_u": 0.1852877839684512,
      "aux_distill/n_active_tok": 175.1875,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5490486391129032,
      "calib/avg_num_step_conf": 2.734375,
      "calib/ece": 0.4747222222222222,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0014238911290322571,
      "calib/mean_conf": 0.01734126984126984,
      "calib/mu_c": 0.018064516129032256,
      "calib/mu_w": 0.016640625,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.00601540706827285,
      "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": 2949.0,
      "completions/max_terminated_length": 2949.0,
      "completions/mean_length": 260.6484375,
      "completions/mean_terminated_length": 260.6484375,
      "completions/min_length": 78.0,
      "completions/min_terminated_length": 78.0,
      "epoch": 0.1024,
      "grad_norm": 0.012914830818772316,
      "learning_rate": 2.888888888888889e-06,
      "loss": 0.0519,
      "num_tokens": 16835490.0,
      "reward": 0.9929591417312622,
      "reward_std": 0.03523903712630272,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.5171683430671692,
      "rewards/format_reward_step": 0.984375,
      "step": 96
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38620558474212885,
      "aux_distill/mean_u": 0.1596223525620612,
      "aux_distill/n_active_tok": 178.3125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5661207622254535,
      "calib/avg_num_step_conf": 2.79296875,
      "calib/ece": 0.2707539682539683,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0015083798882681576,
      "calib/mean_conf": 0.018928571428571427,
      "calib/mu_c": 0.02,
      "calib/mu_w": 0.018491620111731843,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.0,
      "calib/std_conf": 0.005708081949874692,
      "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": 2262.0,
      "completions/max_terminated_length": 2262.0,
      "completions/mean_length": 256.55078125,
      "completions/mean_terminated_length": 257.556884765625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 80.0,
      "epoch": 0.10346666666666667,
      "grad_norm": 0.010874838568270206,
      "learning_rate": 2.861111111111111e-06,
      "loss": 0.159,
      "num_tokens": 17006239.0,
      "reward": 0.985979437828064,
      "reward_std": 0.060727715492248535,
      "rewards/accuracy_reward_step": 0.28515625,
      "rewards/final_brier_reward_step": 0.7063339948654175,
      "rewards/format_reward_step": 0.98046875,
      "step": 97
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3821967989206314,
      "aux_distill/mean_u": 0.23914681162299736,
      "aux_distill/n_active_tok": 170.96875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6189378630967416,
      "calib/avg_num_step_conf": 2.66015625,
      "calib/ece": 0.399921568627451,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0031169487244253544,
      "calib/mean_conf": 0.019686274509803925,
      "calib/mu_c": 0.021495327102803732,
      "calib/mu_w": 0.018378378378378378,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.006316769452560481,
      "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": 2458.0,
      "completions/max_terminated_length": 2458.0,
      "completions/mean_length": 251.21875,
      "completions/mean_terminated_length": 251.21875,
      "completions/min_length": 67.0,
      "completions/min_terminated_length": 67.0,
      "epoch": 0.10453333333333334,
      "grad_norm": 0.014738868921995163,
      "learning_rate": 2.8333333333333335e-06,
      "loss": 0.0794,
      "num_tokens": 17176735.0,
      "reward": 0.9970546960830688,
      "reward_std": 0.040115561336278915,
      "rewards/accuracy_reward_step": 0.41796875,
      "rewards/final_brier_reward_step": 0.5878593921661377,
      "rewards/format_reward_step": 0.98828125,
      "step": 98
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38096280861645937,
      "aux_distill/mean_u": 0.17522439725132843,
      "aux_distill/n_active_tok": 177.0,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5390576242978989,
      "calib/avg_num_step_conf": 2.76953125,
      "calib/ece": 0.1591372549019608,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0013865196588308745,
      "calib/mean_conf": 0.021254901960784313,
      "calib/mu_c": 0.022391304347826088,
      "calib/mu_w": 0.021004784688995214,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.00713962468738118,
      "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": 1888.0,
      "completions/max_terminated_length": 1888.0,
      "completions/mean_length": 266.7265625,
      "completions/mean_terminated_length": 266.7265625,
      "completions/min_length": 75.0,
      "completions/min_terminated_length": 75.0,
      "epoch": 0.1056,
      "grad_norm": 0.010448206216096878,
      "learning_rate": 2.805555555555556e-06,
      "loss": 0.0491,
      "num_tokens": 17350817.0,
      "reward": 0.999866783618927,
      "reward_std": 0.015927515923976898,
      "rewards/accuracy_reward_step": 0.1796875,
      "rewards/final_brier_reward_step": 0.823952317237854,
      "rewards/format_reward_step": 0.99609375,
      "step": 99
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3705893671140075,
      "aux_distill/mean_u": 0.18667120212281332,
      "aux_distill/n_active_tok": 173.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5634003708458137,
      "calib/avg_num_step_conf": 2.6796875,
      "calib/ece": 0.3024505928853755,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0018741976893453145,
      "calib/mean_conf": 0.0216600790513834,
      "calib/mu_c": 0.022926829268292682,
      "calib/mu_w": 0.021052631578947368,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.006683933541162994,
      "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": 2583.0,
      "completions/max_terminated_length": 2583.0,
      "completions/mean_length": 260.68359375,
      "completions/mean_terminated_length": 261.7059020996094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.10666666666666667,
      "grad_norm": 0.012456218712031841,
      "learning_rate": 2.7777777777777783e-06,
      "loss": 0.0995,
      "num_tokens": 17524960.0,
      "reward": 0.9914649724960327,
      "reward_std": 0.0515936017036438,
      "rewards/accuracy_reward_step": 0.3203125,
      "rewards/final_brier_reward_step": 0.678242564201355,
      "rewards/format_reward_step": 0.984375,
      "step": 100
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3266990319825709,
      "aux_distill/mean_u": 0.12328381091635676,
      "aux_distill/n_active_tok": 181.78125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5713571428571429,
      "calib/avg_num_step_conf": 2.8359375,
      "calib/ece": 0.2909411764705882,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.002317857142857146,
      "calib/mean_conf": 0.0227843137254902,
      "calib/mu_c": 0.024375,
      "calib/mu_w": 0.022057142857142855,
      "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": 0.008194483379833082,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2173.0,
      "completions/max_terminated_length": 2173.0,
      "completions/mean_length": 299.1015625,
      "completions/mean_terminated_length": 299.1015625,
      "completions/min_length": 103.0,
      "completions/min_terminated_length": 103.0,
      "epoch": 0.10773333333333333,
      "grad_norm": 0.012489385902881622,
      "learning_rate": 2.7500000000000004e-06,
      "loss": 0.0222,
      "num_tokens": 17708522.0,
      "reward": 1.0034189224243164,
      "reward_std": 0.018446702510118484,
      "rewards/accuracy_reward_step": 0.3125,
      "rewards/final_brier_reward_step": 0.6982441544532776,
      "rewards/format_reward_step": 0.99609375,
      "step": 101
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35586046427488327,
      "aux_distill/mean_u": 0.1803295721689139,
      "aux_distill/n_active_tok": 199.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5229989932041279,
      "calib/avg_num_step_conf": 3.109375,
      "calib/ece": 0.4353754940711463,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0004423609363201561,
      "calib/mean_conf": 0.023122529644268777,
      "calib/mu_c": 0.02336206896551724,
      "calib/mu_w": 0.022919708029197083,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.008108259971591179,
      "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": 1974.0,
      "completions/max_terminated_length": 1974.0,
      "completions/mean_length": 249.71484375,
      "completions/mean_terminated_length": 250.6941375732422,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 82.0,
      "epoch": 0.1088,
      "grad_norm": 0.01360093243420124,
      "learning_rate": 2.7222222222222224e-06,
      "loss": 0.0679,
      "num_tokens": 17879145.0,
      "reward": 0.9985705018043518,
      "reward_std": 0.03953090310096741,
      "rewards/accuracy_reward_step": 0.453125,
      "rewards/final_brier_reward_step": 0.5557347536087036,
      "rewards/format_reward_step": 0.98828125,
      "step": 102
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3942983839660883,
      "aux_distill/mean_u": 0.19361853840302062,
      "aux_distill/n_active_tok": 183.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5702050433010698,
      "calib/avg_num_step_conf": 2.86328125,
      "calib/ece": 0.3844705882352941,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.002423586347427408,
      "calib/mean_conf": 0.023372549019607843,
      "calib/mu_c": 0.02480769230769231,
      "calib/mu_w": 0.0223841059602649,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.007695072995395158,
      "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": 2199.0,
      "completions/max_terminated_length": 2199.0,
      "completions/mean_length": 286.59375,
      "completions/mean_terminated_length": 286.59375,
      "completions/min_length": 92.0,
      "completions/min_terminated_length": 92.0,
      "epoch": 0.10986666666666667,
      "grad_norm": 0.012631945312023163,
      "learning_rate": 2.6944444444444444e-06,
      "loss": 0.0182,
      "num_tokens": 18057065.0,
      "reward": 1.0038397312164307,
      "reward_std": 0.023325443267822266,
      "rewards/accuracy_reward_step": 0.40625,
      "rewards/final_brier_reward_step": 0.6092422008514404,
      "rewards/format_reward_step": 0.9921875,
      "step": 103
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3478834149427712,
      "aux_distill/mean_u": 0.15305093505171327,
      "aux_distill/n_active_tok": 188.34375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5450406276267863,
      "calib/avg_num_step_conf": 2.9453125,
      "calib/ece": 0.29964705882352943,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0016972541328103066,
      "calib/mean_conf": 0.02584313725490196,
      "calib/mu_c": 0.02698795180722891,
      "calib/mu_w": 0.0252906976744186,
      "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": 0.009371154215682412,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/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": 256.73046875,
      "completions/mean_terminated_length": 257.7372741699219,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.11093333333333333,
      "grad_norm": 0.012816665694117546,
      "learning_rate": 2.666666666666667e-06,
      "loss": 0.0176,
      "num_tokens": 18229468.0,
      "reward": 1.0044673681259155,
      "reward_std": 0.019612101837992668,
      "rewards/accuracy_reward_step": 0.32421875,
      "rewards/final_brier_reward_step": 0.688622236251831,
      "rewards/format_reward_step": 0.99609375,
      "step": 104
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35361317032948136,
      "aux_distill/mean_u": 0.14714659951818576,
      "aux_distill/n_active_tok": 180.75,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5229885057471264,
      "calib/avg_num_step_conf": 2.82421875,
      "calib/ece": 0.2880314960629921,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0011982758620689739,
      "calib/mean_conf": 0.026929133858267715,
      "calib/mu_c": 0.027750000000000004,
      "calib/mu_w": 0.02655172413793103,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.010234403230575732,
      "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": 1993.0,
      "completions/max_terminated_length": 1993.0,
      "completions/mean_length": 273.28125,
      "completions/mean_terminated_length": 274.35296630859375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 74.0,
      "epoch": 0.112,
      "grad_norm": 0.012089966796338558,
      "learning_rate": 2.6388888888888893e-06,
      "loss": 0.0422,
      "num_tokens": 18405188.0,
      "reward": 1.0004476308822632,
      "reward_std": 0.03112502582371235,
      "rewards/accuracy_reward_step": 0.3125,
      "rewards/final_brier_reward_step": 0.6962078213691711,
      "rewards/format_reward_step": 0.9921875,
      "step": 105
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3795621353201568,
      "aux_distill/mean_u": 0.18529847398515106,
      "aux_distill/n_active_tok": 176.75,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6127355875831486,
      "calib/avg_num_step_conf": 2.76171875,
      "calib/ece": 0.32047619047619047,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.004046563192904654,
      "calib/mean_conf": 0.02873015873015873,
      "calib/mu_c": 0.031363636363636364,
      "calib/mu_w": 0.02731707317073171,
      "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": 0.010156249697172312,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 1976.0,
      "completions/max_terminated_length": 1976.0,
      "completions/mean_length": 276.05859375,
      "completions/mean_terminated_length": 276.05859375,
      "completions/min_length": 99.0,
      "completions/min_terminated_length": 99.0,
      "epoch": 0.11306666666666666,
      "grad_norm": 0.010598819702863693,
      "learning_rate": 2.6111111111111113e-06,
      "loss": 0.1359,
      "num_tokens": 18580443.0,
      "reward": 0.9904502630233765,
      "reward_std": 0.06372330337762833,
      "rewards/accuracy_reward_step": 0.34375,
      "rewards/final_brier_reward_step": 0.6605879068374634,
      "rewards/format_reward_step": 0.9765625,
      "step": 106
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3647639914415777,
      "aux_distill/mean_u": 0.1525406433967182,
      "aux_distill/n_active_tok": 194.6875,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5104798263343064,
      "calib/avg_num_step_conf": 3.0234375,
      "calib/ece": 0.25667968750000003,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0005980986600793471,
      "calib/mean_conf": 0.028476562499999997,
      "calib/mu_c": 0.028904109589041094,
      "calib/mu_w": 0.028306010928961747,
      "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": 0.010987538995771243,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 838.0,
      "completions/max_terminated_length": 838.0,
      "completions/mean_length": 268.375,
      "completions/mean_terminated_length": 269.4274597167969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 90.0,
      "epoch": 0.11413333333333334,
      "grad_norm": 0.012022835202515125,
      "learning_rate": 2.5833333333333337e-06,
      "loss": 0.0163,
      "num_tokens": 18753763.0,
      "reward": 1.0077763795852661,
      "reward_std": 0.008537918329238892,
      "rewards/accuracy_reward_step": 0.28515625,
      "rewards/final_brier_reward_step": 0.7303965091705322,
      "rewards/format_reward_step": 1.0,
      "step": 107
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33226052671670914,
      "aux_distill/mean_u": 0.19509699005745737,
      "aux_distill/n_active_tok": 203.0,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5082465277777778,
      "calib/avg_num_step_conf": 3.171875,
      "calib/ece": 0.4675196850393701,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0014880952380952397,
      "calib/mean_conf": 0.03051181102362205,
      "calib/mu_c": 0.02976190476190476,
      "calib/mu_w": 0.03125,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.000984251968503937,
      "calib/std_conf": 0.017460130575465912,
      "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": 1999.0,
      "completions/max_terminated_length": 1999.0,
      "completions/mean_length": 288.1484375,
      "completions/mean_terminated_length": 288.1484375,
      "completions/min_length": 93.0,
      "completions/min_terminated_length": 93.0,
      "epoch": 0.1152,
      "grad_norm": 0.010769481770694256,
      "learning_rate": 2.5555555555555557e-06,
      "loss": 0.0623,
      "num_tokens": 18930761.0,
      "reward": 1.0023183822631836,
      "reward_std": 0.02828032523393631,
      "rewards/accuracy_reward_step": 0.4921875,
      "rewards/final_brier_reward_step": 0.5241679549217224,
      "rewards/format_reward_step": 0.98828125,
      "step": 108
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3249229663051665,
      "aux_distill/mean_u": 0.1310028577043218,
      "aux_distill/n_active_tok": 197.28125,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5770042194092827,
      "calib/avg_num_step_conf": 3.08203125,
      "calib/ece": 0.34219123505976096,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0035123179529059495,
      "calib/mean_conf": 0.028326693227091634,
      "calib/mu_c": 0.03053763440860215,
      "calib/mu_w": 0.0270253164556962,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.0,
      "calib/std_conf": 0.010766853953004685,
      "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": 1805.0,
      "completions/max_terminated_length": 1805.0,
      "completions/mean_length": 320.44921875,
      "completions/mean_terminated_length": 321.7059020996094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 98.0,
      "epoch": 0.11626666666666667,
      "grad_norm": 0.009666023775935173,
      "learning_rate": 2.5277777777777778e-06,
      "loss": 0.0355,
      "num_tokens": 19117396.0,
      "reward": 0.9911122918128967,
      "reward_std": 0.045420363545417786,
      "rewards/accuracy_reward_step": 0.36328125,
      "rewards/final_brier_reward_step": 0.6384745836257935,
      "rewards/format_reward_step": 0.98046875,
      "step": 109
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3342225840315223,
      "aux_distill/mean_u": 0.13468734558034612,
      "aux_distill/n_active_tok": 203.1875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5692927170868347,
      "calib/avg_num_step_conf": 3.12109375,
      "calib/ece": 0.2990551181102362,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0030434173669467712,
      "calib/mean_conf": 0.03165354330708662,
      "calib/mu_c": 0.03369047619047619,
      "calib/mu_w": 0.030647058823529416,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.010850725112934664,
      "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": 1881.0,
      "completions/max_terminated_length": 1881.0,
      "completions/mean_length": 290.29296875,
      "completions/mean_terminated_length": 291.431396484375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 89.0,
      "epoch": 0.11733333333333333,
      "grad_norm": 0.010616586543619633,
      "learning_rate": 2.5e-06,
      "loss": 0.0857,
      "num_tokens": 19296631.0,
      "reward": 1.000656247138977,
      "reward_std": 0.035624802112579346,
      "rewards/accuracy_reward_step": 0.328125,
      "rewards/final_brier_reward_step": 0.6849062442779541,
      "rewards/format_reward_step": 0.98828125,
      "step": 110
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36175938649103045,
      "aux_distill/mean_u": 0.17420325036256049,
      "aux_distill/n_active_tok": 215.34375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5661924900679226,
      "calib/avg_num_step_conf": 3.36328125,
      "calib/ece": 0.36768627450980396,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0026797385620914924,
      "calib/mean_conf": 0.03231372549019607,
      "calib/mu_c": 0.03392156862745097,
      "calib/mu_w": 0.03124183006535948,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.011467755641539287,
      "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": 1971.0,
      "completions/max_terminated_length": 1971.0,
      "completions/mean_length": 309.6328125,
      "completions/mean_terminated_length": 309.6328125,
      "completions/min_length": 100.0,
      "completions/min_terminated_length": 100.0,
      "epoch": 0.1184,
      "grad_norm": 0.010419880039989948,
      "learning_rate": 2.4722222222222226e-06,
      "loss": 0.074,
      "num_tokens": 19483305.0,
      "reward": 1.0051193237304688,
      "reward_std": 0.03306365758180618,
      "rewards/accuracy_reward_step": 0.3984375,
      "rewards/final_brier_reward_step": 0.6196136474609375,
      "rewards/format_reward_step": 0.9921875,
      "step": 111
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3407304754946381,
      "aux_distill/mean_u": 0.20758217514789062,
      "aux_distill/n_active_tok": 210.5,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5607392090979582,
      "calib/avg_num_step_conf": 3.29296875,
      "calib/ece": 0.38861111111111113,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0025316619281468035,
      "calib/mean_conf": 0.032023809523809524,
      "calib/mu_c": 0.03349056603773585,
      "calib/mu_w": 0.030958904109589045,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98046875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.010440279359604128,
      "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": 2222.0,
      "completions/max_terminated_length": 2222.0,
      "completions/mean_length": 341.10546875,
      "completions/mean_terminated_length": 342.4431457519531,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 99.0,
      "epoch": 0.11946666666666667,
      "grad_norm": 0.01065851654857397,
      "learning_rate": 2.4444444444444447e-06,
      "loss": 0.0805,
      "num_tokens": 19678548.0,
      "reward": 0.9859668016433716,
      "reward_std": 0.05308392643928528,
      "rewards/accuracy_reward_step": 0.4140625,
      "rewards/final_brier_reward_step": 0.5852148532867432,
      "rewards/format_reward_step": 0.97265625,
      "step": 112
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31915219803340733,
      "aux_distill/mean_u": 0.13680159858362684,
      "aux_distill/n_active_tok": 241.75,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5020853698273052,
      "calib/avg_num_step_conf": 3.77734375,
      "calib/ece": 0.3543307086614173,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 3.5190615835761774e-05,
      "calib/mean_conf": 0.03543307086614173,
      "calib/mu_c": 0.03545454545454544,
      "calib/mu_w": 0.03541935483870968,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.011722238473165,
      "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": 2200.0,
      "completions/max_terminated_length": 2200.0,
      "completions/mean_length": 309.89453125,
      "completions/mean_terminated_length": 309.89453125,
      "completions/min_length": 117.0,
      "completions/min_terminated_length": 117.0,
      "epoch": 0.12053333333333334,
      "grad_norm": 0.012033089064061642,
      "learning_rate": 2.4166666666666667e-06,
      "loss": 0.0769,
      "num_tokens": 19863081.0,
      "reward": 1.0052074193954468,
      "reward_std": 0.03318488597869873,
      "rewards/accuracy_reward_step": 0.38671875,
      "rewards/final_brier_reward_step": 0.6315085887908936,
      "rewards/format_reward_step": 0.9921875,
      "step": 113
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31962078250944614,
      "aux_distill/mean_u": 0.14795968757927555,
      "aux_distill/n_active_tok": 236.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.568682344158268,
      "calib/avg_num_step_conf": 3.69140625,
      "calib/ece": 0.4123137254901961,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0026819708846584564,
      "calib/mean_conf": 0.03474509803921568,
      "calib/mu_c": 0.036228070175438595,
      "calib/mu_w": 0.03354609929078014,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.012296728621642014,
      "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": 1992.0,
      "completions/max_terminated_length": 1992.0,
      "completions/mean_length": 299.78125,
      "completions/mean_terminated_length": 299.78125,
      "completions/min_length": 89.0,
      "completions/min_terminated_length": 89.0,
      "epoch": 0.1216,
      "grad_norm": 0.012581607326865196,
      "learning_rate": 2.388888888888889e-06,
      "loss": 0.0884,
      "num_tokens": 20044849.0,
      "reward": 1.011549949645996,
      "reward_std": 0.023771952837705612,
      "rewards/accuracy_reward_step": 0.4453125,
      "rewards/final_brier_reward_step": 0.5816937685012817,
      "rewards/format_reward_step": 0.99609375,
      "step": 114
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31680090888403356,
      "aux_distill/mean_u": 0.13020203300884048,
      "aux_distill/n_active_tok": 258.5,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5182643427741467,
      "calib/avg_num_step_conf": 4.0390625,
      "calib/ece": 0.28820717131474105,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 9.731299927378395e-05,
      "calib/mean_conf": 0.034501992031872514,
      "calib/mu_c": 0.0345679012345679,
      "calib/mu_w": 0.034470588235294114,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0,
      "calib/std_conf": 0.013090665906683839,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 349.6015625,
      "completions/mean_terminated_length": 349.6015625,
      "completions/min_length": 85.0,
      "completions/min_terminated_length": 85.0,
      "epoch": 0.12266666666666666,
      "grad_norm": 0.011446863412857056,
      "learning_rate": 2.361111111111111e-06,
      "loss": 0.1378,
      "num_tokens": 20239611.0,
      "reward": 0.9907386302947998,
      "reward_std": 0.06587958335876465,
      "rewards/accuracy_reward_step": 0.31640625,
      "rewards/final_brier_reward_step": 0.6846023201942444,
      "rewards/format_reward_step": 0.98046875,
      "step": 115
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3188343502115458,
      "aux_distill/mean_u": 0.125653797523076,
      "aux_distill/n_active_tok": 264.75,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.4851329131268841,
      "calib/avg_num_step_conf": 4.13671875,
      "calib/ece": 0.3164426877470356,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.000258289942449981,
      "calib/mean_conf": 0.035335968379446636,
      "calib/mu_c": 0.0351685393258427,
      "calib/mu_w": 0.03542682926829268,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.012620399241968066,
      "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": 2015.0,
      "completions/max_terminated_length": 2015.0,
      "completions/mean_length": 370.484375,
      "completions/mean_terminated_length": 370.484375,
      "completions/min_length": 73.0,
      "completions/min_terminated_length": 73.0,
      "epoch": 0.12373333333333333,
      "grad_norm": 0.01038755476474762,
      "learning_rate": 2.3333333333333336e-06,
      "loss": 0.0656,
      "num_tokens": 20438975.0,
      "reward": 0.999812126159668,
      "reward_std": 0.04449833184480667,
      "rewards/accuracy_reward_step": 0.34765625,
      "rewards/final_brier_reward_step": 0.6636867523193359,
      "rewards/format_reward_step": 0.98828125,
      "step": 116
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2782537774182856,
      "aux_distill/mean_u": 0.11942123560487757,
      "aux_distill/n_active_tok": 270.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5806763658876334,
      "calib/avg_num_step_conf": 4.21875,
      "calib/ece": 0.245098814229249,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0030761492029097656,
      "calib/mean_conf": 0.035533596837944666,
      "calib/mu_c": 0.03774647887323943,
      "calib/mu_w": 0.034670329670329667,
      "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": 0.01239229717829777,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2360.0,
      "completions/max_terminated_length": 2360.0,
      "completions/mean_length": 394.31640625,
      "completions/mean_terminated_length": 394.31640625,
      "completions/min_length": 89.0,
      "completions/min_terminated_length": 89.0,
      "epoch": 0.1248,
      "grad_norm": 0.008689876645803452,
      "learning_rate": 2.305555555555556e-06,
      "loss": 0.0694,
      "num_tokens": 20646520.0,
      "reward": 0.9980502128601074,
      "reward_std": 0.04275971278548241,
      "rewards/accuracy_reward_step": 0.27734375,
      "rewards/final_brier_reward_step": 0.7304753661155701,
      "rewards/format_reward_step": 0.98828125,
      "step": 117
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.2848955246154219,
      "aux_distill/mean_u": 0.13537754193122786,
      "aux_distill/n_active_tok": 264.875,
      "calib/answer_extract_rate": 0.9765625,
      "calib/auroc": 0.4693205529473136,
      "calib/avg_num_step_conf": 4.140625,
      "calib/ece": 0.39548,
      "calib/final_conf_rate": 0.9765625,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.001208920187793422,
      "calib/mean_conf": 0.03652,
      "calib/mu_c": 0.03583333333333333,
      "calib/mu_w": 0.03704225352112675,
      "calib/nonempty_final_conf_rate": 0.9765625,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.012340567247902343,
      "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": 2594.0,
      "completions/max_terminated_length": 2594.0,
      "completions/mean_length": 373.71484375,
      "completions/mean_terminated_length": 375.180419921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.12586666666666665,
      "grad_norm": 0.010260322131216526,
      "learning_rate": 2.277777777777778e-06,
      "loss": 0.1309,
      "num_tokens": 20846199.0,
      "reward": 0.9909540414810181,
      "reward_std": 0.052009761333465576,
      "rewards/accuracy_reward_step": 0.421875,
      "rewards/final_brier_reward_step": 0.5834707021713257,
      "rewards/format_reward_step": 0.9765625,
      "step": 118
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32525242306292057,
      "aux_distill/mean_u": 0.17335328483206366,
      "aux_distill/n_active_tok": 251.25,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5693690320164763,
      "calib/avg_num_step_conf": 3.94140625,
      "calib/ece": 0.389609375,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.002033327092304804,
      "calib/mean_conf": 0.036171875000000006,
      "calib/mu_c": 0.037339449541284396,
      "calib/mu_w": 0.03530612244897959,
      "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": 0.011734104524179721,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1056.0,
      "completions/max_terminated_length": 1056.0,
      "completions/mean_length": 348.234375,
      "completions/mean_terminated_length": 349.6000061035156,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 102.0,
      "epoch": 0.12693333333333334,
      "grad_norm": 0.009621622040867805,
      "learning_rate": 2.25e-06,
      "loss": 0.0131,
      "num_tokens": 21040411.0,
      "reward": 1.0151753425598145,
      "reward_std": 0.012086945585906506,
      "rewards/accuracy_reward_step": 0.42578125,
      "rewards/final_brier_reward_step": 0.6045695543289185,
      "rewards/format_reward_step": 1.0,
      "step": 119
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32881034817546606,
      "aux_distill/mean_u": 0.1642785781487071,
      "aux_distill/n_active_tok": 276.15625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5925612298071912,
      "calib/avg_num_step_conf": 4.30859375,
      "calib/ece": 0.3640711462450593,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.004301719645648773,
      "calib/mean_conf": 0.03513833992094862,
      "calib/mu_c": 0.03772277227722772,
      "calib/mu_w": 0.03342105263157895,
      "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": 0.011948532266813949,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2931.0,
      "completions/max_terminated_length": 2931.0,
      "completions/mean_length": 360.81640625,
      "completions/mean_terminated_length": 360.81640625,
      "completions/min_length": 77.0,
      "completions/min_terminated_length": 77.0,
      "epoch": 0.128,
      "grad_norm": 0.009301810525357723,
      "learning_rate": 2.222222222222222e-06,
      "loss": 0.1065,
      "num_tokens": 21239468.0,
      "reward": 1.0024833679199219,
      "reward_std": 0.044663794338703156,
      "rewards/accuracy_reward_step": 0.39453125,
      "rewards/final_brier_reward_step": 0.6221542954444885,
      "rewards/format_reward_step": 0.98828125,
      "step": 120
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32076170202344656,
      "aux_distill/mean_u": 0.1362645584935083,
      "aux_distill/n_active_tok": 259.75,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.532148524278007,
      "calib/avg_num_step_conf": 4.05859375,
      "calib/ece": 0.42178571428571426,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0010428435417327853,
      "calib/mean_conf": 0.03456349206349207,
      "calib/mu_c": 0.035130434782608695,
      "calib/mu_w": 0.03408759124087591,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.012416704079039853,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2775.0,
      "completions/max_terminated_length": 2775.0,
      "completions/mean_length": 385.90625,
      "completions/mean_terminated_length": 387.41961669921875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 134.0,
      "epoch": 0.12906666666666666,
      "grad_norm": 0.009882324375212193,
      "learning_rate": 2.1944444444444445e-06,
      "loss": 0.1093,
      "num_tokens": 21443316.0,
      "reward": 0.9994924068450928,
      "reward_std": 0.05566861853003502,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.5653910040855408,
      "rewards/format_reward_step": 0.984375,
      "step": 121
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36098944093100727,
      "aux_distill/mean_u": 0.16837954315821466,
      "aux_distill/n_active_tok": 266.3125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6004809567138958,
      "calib/avg_num_step_conf": 4.08984375,
      "calib/ece": 0.34929411764705887,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0037800597946184844,
      "calib/mean_conf": 0.035019607843137256,
      "calib/mu_c": 0.0373469387755102,
      "calib/mu_w": 0.033566878980891716,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.011371602394679636,
      "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": 2043.0,
      "completions/max_terminated_length": 2043.0,
      "completions/mean_length": 338.0546875,
      "completions/mean_terminated_length": 338.0546875,
      "completions/min_length": 131.0,
      "completions/min_terminated_length": 131.0,
      "epoch": 0.13013333333333332,
      "grad_norm": 0.010515622794628143,
      "learning_rate": 2.166666666666667e-06,
      "loss": 0.0752,
      "num_tokens": 21637202.0,
      "reward": 1.0097154378890991,
      "reward_std": 0.022465920075774193,
      "rewards/accuracy_reward_step": 0.3828125,
      "rewards/final_brier_reward_step": 0.6405245661735535,
      "rewards/format_reward_step": 0.99609375,
      "step": 122
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3363469988107681,
      "aux_distill/mean_u": 0.1908947539477813,
      "aux_distill/n_active_tok": 285.21875,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.6363573311138854,
      "calib/avg_num_step_conf": 4.421875,
      "calib/ece": 0.32095617529880477,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.005334304341794979,
      "calib/mean_conf": 0.034422310756972115,
      "calib/mu_c": 0.037865168539325846,
      "calib/mu_w": 0.032530864197530866,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.00039840637450199205,
      "calib/std_conf": 0.012810484230277445,
      "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": 2173.0,
      "completions/max_terminated_length": 2173.0,
      "completions/mean_length": 426.55859375,
      "completions/mean_terminated_length": 426.55859375,
      "completions/min_length": 145.0,
      "completions/min_terminated_length": 145.0,
      "epoch": 0.1312,
      "grad_norm": 0.009776564314961433,
      "learning_rate": 2.138888888888889e-06,
      "loss": 0.0799,
      "num_tokens": 21851689.0,
      "reward": 0.9929714798927307,
      "reward_std": 0.049630723893642426,
      "rewards/accuracy_reward_step": 0.34765625,
      "rewards/final_brier_reward_step": 0.6578179597854614,
      "rewards/format_reward_step": 0.98046875,
      "step": 123
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3255585916340351,
      "aux_distill/mean_u": 0.13354474605010633,
      "aux_distill/n_active_tok": 275.28125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5778886587585514,
      "calib/avg_num_step_conf": 4.30078125,
      "calib/ece": 0.4089763779527559,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.003709282620975335,
      "calib/mean_conf": 0.03590551181102362,
      "calib/mu_c": 0.0379646017699115,
      "calib/mu_w": 0.03425531914893617,
      "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": 0.011525687210875166,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2292.0,
      "completions/max_terminated_length": 2292.0,
      "completions/mean_length": 345.0234375,
      "completions/mean_terminated_length": 345.0234375,
      "completions/min_length": 111.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.13226666666666667,
      "grad_norm": 0.010006045922636986,
      "learning_rate": 2.1111111111111114e-06,
      "loss": 0.0613,
      "num_tokens": 22046831.0,
      "reward": 1.0082398653030396,
      "reward_std": 0.03543369472026825,
      "rewards/accuracy_reward_step": 0.44140625,
      "rewards/final_brier_reward_step": 0.5828859210014343,
      "rewards/format_reward_step": 0.9921875,
      "step": 124
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.29937303997576237,
      "aux_distill/mean_u": 0.15761940309284003,
      "aux_distill/n_active_tok": 282.5,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.4592781750085121,
      "calib/avg_num_step_conf": 4.41796875,
      "calib/ece": 0.3146062992125984,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.0025093632958801404,
      "calib/mean_conf": 0.035787401574803146,
      "calib/mu_c": 0.03415730337078652,
      "calib/mu_w": 0.03666666666666666,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.013070451089131675,
      "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": 2252.0,
      "completions/max_terminated_length": 2252.0,
      "completions/mean_length": 364.56640625,
      "completions/mean_terminated_length": 365.99609375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 126.0,
      "epoch": 0.13333333333333333,
      "grad_norm": 0.01030960027128458,
      "learning_rate": 2.0833333333333334e-06,
      "loss": 0.0573,
      "num_tokens": 22244968.0,
      "reward": 1.0033423900604248,
      "reward_std": 0.03312192112207413,
      "rewards/accuracy_reward_step": 0.34765625,
      "rewards/final_brier_reward_step": 0.6668410301208496,
      "rewards/format_reward_step": 0.9921875,
      "step": 125
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3546295613050461,
      "aux_distill/mean_u": 0.19248182451684565,
      "aux_distill/n_active_tok": 274.875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5799744897959184,
      "calib/avg_num_step_conf": 4.29296875,
      "calib/ece": 0.4076587301587301,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.002410714285714294,
      "calib/mean_conf": 0.03678571428571428,
      "calib/mu_c": 0.038125000000000006,
      "calib/mu_w": 0.03571428571428571,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.01292475608001525,
      "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": 2221.0,
      "completions/max_terminated_length": 2221.0,
      "completions/mean_length": 373.27734375,
      "completions/mean_terminated_length": 373.27734375,
      "completions/min_length": 95.0,
      "completions/min_terminated_length": 95.0,
      "epoch": 0.1344,
      "grad_norm": 0.009459101594984531,
      "learning_rate": 2.0555555555555555e-06,
      "loss": 0.1564,
      "num_tokens": 22445991.0,
      "reward": 1.0003063678741455,
      "reward_std": 0.054473064839839935,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/final_brier_reward_step": 0.5787379145622253,
      "rewards/format_reward_step": 0.984375,
      "step": 126
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3222575797699392,
      "aux_distill/mean_u": 0.15707868353003399,
      "aux_distill/n_active_tok": 285.65625,
      "calib/answer_extract_rate": 0.97265625,
      "calib/auroc": 0.7024915824915825,
      "calib/avg_num_step_conf": 4.453125,
      "calib/ece": 0.36180722891566264,
      "calib/final_conf_rate": 0.97265625,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.009347474747474757,
      "calib/mean_conf": 0.035783132530120484,
      "calib/mu_c": 0.04141414141414142,
      "calib/mu_w": 0.03206666666666666,
      "calib/nonempty_final_conf_rate": 0.97265625,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.013243516130715595,
      "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": 2538.0,
      "completions/max_terminated_length": 2538.0,
      "completions/mean_length": 399.76171875,
      "completions/mean_terminated_length": 399.76171875,
      "completions/min_length": 116.0,
      "completions/min_terminated_length": 116.0,
      "epoch": 0.13546666666666668,
      "grad_norm": 0.009776722639799118,
      "learning_rate": 2.027777777777778e-06,
      "loss": 0.1557,
      "num_tokens": 22652002.0,
      "reward": 0.987963855266571,
      "reward_std": 0.07154573500156403,
      "rewards/accuracy_reward_step": 0.38671875,
      "rewards/final_brier_reward_step": 0.6165527105331421,
      "rewards/format_reward_step": 0.97265625,
      "step": 127
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35741139482706785,
      "aux_distill/mean_u": 0.1896862522924126,
      "aux_distill/n_active_tok": 248.8125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.549607843137255,
      "calib/avg_num_step_conf": 3.86328125,
      "calib/ece": 0.36051383399209486,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0022437908496731954,
      "calib/mean_conf": 0.034743083003952575,
      "calib/mu_c": 0.03609999999999999,
      "calib/mu_w": 0.0338562091503268,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.012527960884909475,
      "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": 2875.0,
      "completions/max_terminated_length": 2875.0,
      "completions/mean_length": 395.7734375,
      "completions/mean_terminated_length": 395.7734375,
      "completions/min_length": 136.0,
      "completions/min_terminated_length": 136.0,
      "epoch": 0.13653333333333334,
      "grad_norm": 0.0100556043908,
      "learning_rate": 2.0000000000000003e-06,
      "loss": 0.1223,
      "num_tokens": 22859984.0,
      "reward": 0.9996402263641357,
      "reward_std": 0.05114775896072388,
      "rewards/accuracy_reward_step": 0.390625,
      "rewards/final_brier_reward_step": 0.6242804527282715,
      "rewards/format_reward_step": 0.984375,
      "step": 128
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3307932415045798,
      "aux_distill/mean_u": 0.16508309298781187,
      "aux_distill/n_active_tok": 257.5,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5442307692307692,
      "calib/avg_num_step_conf": 4.02734375,
      "calib/ece": 0.3728346456692913,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0018256410256410338,
      "calib/mean_conf": 0.03661417322834646,
      "calib/mu_c": 0.03769230769230769,
      "calib/mu_w": 0.03586666666666666,
      "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": 0.013673381948982037,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2891.0,
      "completions/max_terminated_length": 2891.0,
      "completions/mean_length": 368.37109375,
      "completions/mean_terminated_length": 368.37109375,
      "completions/min_length": 105.0,
      "completions/min_terminated_length": 105.0,
      "epoch": 0.1376,
      "grad_norm": 0.009598263539373875,
      "learning_rate": 1.9722222222222224e-06,
      "loss": 0.0814,
      "num_tokens": 23056671.0,
      "reward": 1.006742238998413,
      "reward_std": 0.03526865690946579,
      "rewards/accuracy_reward_step": 0.40625,
      "rewards/final_brier_reward_step": 0.6150468587875366,
      "rewards/format_reward_step": 0.9921875,
      "step": 129
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33436241559684277,
      "aux_distill/mean_u": 0.16774004768920198,
      "aux_distill/n_active_tok": 255.0,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6012166060454033,
      "calib/avg_num_step_conf": 3.99609375,
      "calib/ece": 0.4934387351778657,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.004576696350181854,
      "calib/mean_conf": 0.036205533596837945,
      "calib/mu_c": 0.03835820895522387,
      "calib/mu_w": 0.033781512605042016,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.013211163122124604,
      "calib/step_conf_rate": 0.9921875,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.01171875,
      "completions/max_length": 1853.0,
      "completions/max_terminated_length": 1853.0,
      "completions/mean_length": 350.28125,
      "completions/mean_terminated_length": 354.4347839355469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 121.0,
      "epoch": 0.13866666666666666,
      "grad_norm": 0.010324436239898205,
      "learning_rate": 1.944444444444445e-06,
      "loss": 0.0118,
      "num_tokens": 23251631.0,
      "reward": 1.0076253414154053,
      "reward_std": 0.04466870054602623,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/final_brier_reward_step": 0.5035320520401001,
      "rewards/format_reward_step": 0.98828125,
      "step": 130
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32768950890749693,
      "aux_distill/mean_u": 0.1545797506967805,
      "aux_distill/n_active_tok": 249.875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.526496010638298,
      "calib/avg_num_step_conf": 3.90234375,
      "calib/ece": 0.3321259842519685,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0012234042553191452,
      "calib/mean_conf": 0.037952755905511816,
      "calib/mu_c": 0.038723404255319144,
      "calib/mu_w": 0.0375,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.012971111613869993,
      "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": 2391.0,
      "completions/max_terminated_length": 2391.0,
      "completions/mean_length": 368.4921875,
      "completions/mean_terminated_length": 368.4921875,
      "completions/min_length": 111.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.13973333333333332,
      "grad_norm": 0.0099790683016181,
      "learning_rate": 1.916666666666667e-06,
      "loss": 0.0965,
      "num_tokens": 23452173.0,
      "reward": 1.0035020112991333,
      "reward_std": 0.03792167454957962,
      "rewards/accuracy_reward_step": 0.3671875,
      "rewards/final_brier_reward_step": 0.651535153388977,
      "rewards/format_reward_step": 0.98828125,
      "step": 131
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32731077587231994,
      "aux_distill/mean_u": 0.22404103030817973,
      "aux_distill/n_active_tok": 251.3125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6355338834708676,
      "calib/avg_num_step_conf": 3.875,
      "calib/ece": 0.4702371541501977,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.007210552638159548,
      "calib/mean_conf": 0.03964426877470356,
      "calib/mu_c": 0.043178294573643416,
      "calib/mu_w": 0.03596774193548387,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.014645748913053333,
      "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": 1385.0,
      "completions/max_terminated_length": 1385.0,
      "completions/mean_length": 348.46875,
      "completions/mean_terminated_length": 349.8353271484375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 91.0,
      "epoch": 0.1408,
      "grad_norm": 0.009376777336001396,
      "learning_rate": 1.888888888888889e-06,
      "loss": 0.0607,
      "num_tokens": 23646973.0,
      "reward": 1.0091564655303955,
      "reward_std": 0.04505951702594757,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.526125431060791,
      "rewards/format_reward_step": 0.98828125,
      "step": 132
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38718144688755274,
      "aux_distill/mean_u": 0.1867236683741768,
      "aux_distill/n_active_tok": 259.59375,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.6159521590556073,
      "calib/avg_num_step_conf": 4.0546875,
      "calib/ece": 0.2611290322580645,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.004100652376514448,
      "calib/mean_conf": 0.03725806451612904,
      "calib/mu_c": 0.040135135135135135,
      "calib/mu_w": 0.03603448275862069,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.0,
      "calib/std_conf": 0.013550307083618207,
      "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": 2897.0,
      "completions/max_terminated_length": 2897.0,
      "completions/mean_length": 419.609375,
      "completions/mean_terminated_length": 421.2549133300781,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 120.0,
      "epoch": 0.14186666666666667,
      "grad_norm": 0.008629664778709412,
      "learning_rate": 1.8611111111111113e-06,
      "loss": 0.1469,
      "num_tokens": 23860737.0,
      "reward": 0.9756871461868286,
      "reward_std": 0.09578809142112732,
      "rewards/accuracy_reward_step": 0.2890625,
      "rewards/final_brier_reward_step": 0.6974679231643677,
      "rewards/format_reward_step": 0.96484375,
      "step": 133
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3483828450553119,
      "aux_distill/mean_u": 0.17435870078236385,
      "aux_distill/n_active_tok": 245.0625,
      "calib/answer_extract_rate": 0.96875,
      "calib/auroc": 0.5869265389082462,
      "calib/avg_num_step_conf": 3.82421875,
      "calib/ece": 0.30072580645161284,
      "calib/final_conf_rate": 0.96875,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.004849012775842043,
      "calib/mean_conf": 0.037983870967741945,
      "calib/mu_c": 0.04119047619047619,
      "calib/mu_w": 0.036341463414634144,
      "calib/nonempty_final_conf_rate": 0.96875,
      "calib/nonempty_reasoning_rate": 0.98828125,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.014890993094424206,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0078125,
      "completions/max_length": 1958.0,
      "completions/max_terminated_length": 1958.0,
      "completions/mean_length": 400.36328125,
      "completions/mean_terminated_length": 403.5157470703125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.14293333333333333,
      "grad_norm": 0.007807637099176645,
      "learning_rate": 1.8333333333333333e-06,
      "loss": 0.0699,
      "num_tokens": 24072182.0,
      "reward": 0.9793531894683838,
      "reward_std": 0.0878257304430008,
      "rewards/accuracy_reward_step": 0.328125,
      "rewards/final_brier_reward_step": 0.665737509727478,
      "rewards/format_reward_step": 0.96484375,
      "step": 134
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3719960623420775,
      "aux_distill/mean_u": 0.14194528905229298,
      "aux_distill/n_active_tok": 271.46875,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5835403726708074,
      "calib/avg_num_step_conf": 4.234375,
      "calib/ece": 0.31928286852589643,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0038895790200137934,
      "calib/mean_conf": 0.039282868525896415,
      "calib/mu_c": 0.04177777777777777,
      "calib/mu_w": 0.037888198757763975,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.014866057536278997,
      "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": 2971.0,
      "completions/max_terminated_length": 2971.0,
      "completions/mean_length": 390.16015625,
      "completions/mean_terminated_length": 390.16015625,
      "completions/min_length": 104.0,
      "completions/min_terminated_length": 104.0,
      "epoch": 0.144,
      "grad_norm": 0.008397375233471394,
      "learning_rate": 1.8055555555555557e-06,
      "loss": 0.1209,
      "num_tokens": 24277943.0,
      "reward": 0.994291365146637,
      "reward_std": 0.0661432147026062,
      "rewards/accuracy_reward_step": 0.3515625,
      "rewards/final_brier_reward_step": 0.6565515398979187,
      "rewards/format_reward_step": 0.98046875,
      "step": 135
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34026037296280265,
      "aux_distill/mean_u": 0.16830172599808957,
      "aux_distill/n_active_tok": 248.5625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6767696267696268,
      "calib/avg_num_step_conf": 3.88671875,
      "calib/ece": 0.3774308300395257,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.008354568854568853,
      "calib/mean_conf": 0.03758893280632412,
      "calib/mu_c": 0.04247619047619047,
      "calib/mu_w": 0.03412162162162162,
      "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": 0.013131931485992352,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1928.0,
      "completions/max_terminated_length": 1928.0,
      "completions/mean_length": 377.73828125,
      "completions/mean_terminated_length": 379.2196350097656,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 91.0,
      "epoch": 0.14506666666666668,
      "grad_norm": 0.009265841916203499,
      "learning_rate": 1.777777777777778e-06,
      "loss": 0.0783,
      "num_tokens": 24483132.0,
      "reward": 1.0049197673797607,
      "reward_std": 0.044206224381923676,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.6114019155502319,
      "rewards/format_reward_step": 0.98828125,
      "step": 136
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3602973301894963,
      "aux_distill/mean_u": 0.1981645887058841,
      "aux_distill/n_active_tok": 265.6875,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5256769374416433,
      "calib/avg_num_step_conf": 4.1484375,
      "calib/ece": 0.35139442231075696,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.00011938108576765166,
      "calib/mean_conf": 0.041195219123505974,
      "calib/mu_c": 0.04112244897959183,
      "calib/mu_w": 0.04124183006535948,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0010756972111553786,
      "calib/std_conf": 0.01908686729245096,
      "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": 2672.0,
      "completions/max_terminated_length": 2672.0,
      "completions/mean_length": 383.1171875,
      "completions/mean_terminated_length": 383.1171875,
      "completions/min_length": 94.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.14613333333333334,
      "grad_norm": 0.00766771798953414,
      "learning_rate": 1.75e-06,
      "loss": 0.1229,
      "num_tokens": 24688194.0,
      "reward": 0.9952003955841064,
      "reward_std": 0.06639545410871506,
      "rewards/accuracy_reward_step": 0.3828125,
      "rewards/final_brier_reward_step": 0.6271195411682129,
      "rewards/format_reward_step": 0.98046875,
      "step": 137
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.32277982542291284,
      "aux_distill/mean_u": 0.1438092188570127,
      "aux_distill/n_active_tok": 256.09375,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5729961832061068,
      "calib/avg_num_step_conf": 4.0,
      "calib/ece": 0.4808764940239044,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.003192111959287529,
      "calib/mean_conf": 0.04183266932270916,
      "calib/mu_c": 0.0433587786259542,
      "calib/mu_w": 0.04016666666666667,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.00039840637450199205,
      "calib/std_conf": 0.013822576963452685,
      "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": 2305.0,
      "completions/max_terminated_length": 2305.0,
      "completions/mean_length": 361.82421875,
      "completions/mean_terminated_length": 361.82421875,
      "completions/min_length": 105.0,
      "completions/min_terminated_length": 105.0,
      "epoch": 0.1472,
      "grad_norm": 0.01024717278778553,
      "learning_rate": 1.7222222222222224e-06,
      "loss": 0.1336,
      "num_tokens": 24885157.0,
      "reward": 1.0017046928405762,
      "reward_std": 0.07045163214206696,
      "rewards/accuracy_reward_step": 0.51171875,
      "rewards/final_brier_reward_step": 0.5112218856811523,
      "rewards/format_reward_step": 0.98046875,
      "step": 138
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3387621305882931,
      "aux_distill/mean_u": 0.16340000411789907,
      "aux_distill/n_active_tok": 224.0625,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5776785714285715,
      "calib/avg_num_step_conf": 3.50390625,
      "calib/ece": 0.39956349206349207,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00439285714285715,
      "calib/mean_conf": 0.044880952380952376,
      "calib/mu_c": 0.04732142857142858,
      "calib/mu_w": 0.04292857142857143,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.0,
      "calib/std_conf": 0.013077368760425208,
      "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": 1891.0,
      "completions/max_terminated_length": 1891.0,
      "completions/mean_length": 332.26171875,
      "completions/mean_terminated_length": 332.26171875,
      "completions/min_length": 118.0,
      "completions/min_terminated_length": 118.0,
      "epoch": 0.14826666666666666,
      "grad_norm": 0.009077508933842182,
      "learning_rate": 1.6944444444444446e-06,
      "loss": 0.0794,
      "num_tokens": 25073312.0,
      "reward": 1.0000994205474854,
      "reward_std": 0.061380185186862946,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/final_brier_reward_step": 0.5822300910949707,
      "rewards/format_reward_step": 0.98046875,
      "step": 139
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3609836152754724,
      "aux_distill/mean_u": 0.14727386832180747,
      "aux_distill/n_active_tok": 233.96875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6066162570888468,
      "calib/avg_num_step_conf": 3.6328125,
      "calib/ece": 0.5015019762845849,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00501449275362318,
      "calib/mean_conf": 0.04395256916996048,
      "calib/mu_c": 0.04623188405797101,
      "calib/mu_w": 0.04121739130434783,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.0,
      "calib/std_conf": 0.013520213359584744,
      "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": 1732.0,
      "completions/max_terminated_length": 1732.0,
      "completions/mean_length": 334.64453125,
      "completions/mean_terminated_length": 335.9568786621094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 110.0,
      "epoch": 0.14933333333333335,
      "grad_norm": 0.009565300308167934,
      "learning_rate": 1.6666666666666667e-06,
      "loss": 0.0898,
      "num_tokens": 25263997.0,
      "reward": 1.0043518543243408,
      "reward_std": 0.06932803243398666,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/final_brier_reward_step": 0.48917266726493835,
      "rewards/format_reward_step": 0.98046875,
      "step": 140
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34659315133467317,
      "aux_distill/mean_u": 0.1644295336444299,
      "aux_distill/n_active_tok": 262.09375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6452615384615384,
      "calib/avg_num_step_conf": 4.05859375,
      "calib/ece": 0.44654901960784316,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.007910769230769235,
      "calib/mean_conf": 0.043647058823529414,
      "calib/mu_c": 0.04768000000000001,
      "calib/mu_w": 0.03976923076923077,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.01504349710256798,
      "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": 2254.0,
      "completions/max_terminated_length": 2254.0,
      "completions/mean_length": 357.1796875,
      "completions/mean_terminated_length": 357.1796875,
      "completions/min_length": 122.0,
      "completions/min_terminated_length": 122.0,
      "epoch": 0.1504,
      "grad_norm": 0.008549941703677177,
      "learning_rate": 1.638888888888889e-06,
      "loss": 0.0807,
      "num_tokens": 25462531.0,
      "reward": 1.014409065246582,
      "reward_std": 0.036273662000894547,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/final_brier_reward_step": 0.5483492016792297,
      "rewards/format_reward_step": 0.9921875,
      "step": 141
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3710846067406237,
      "aux_distill/mean_u": 0.1729861332487001,
      "aux_distill/n_active_tok": 231.875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5972666830829846,
      "calib/avg_num_step_conf": 3.63671875,
      "calib/ece": 0.4383529411764706,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0051446688007879965,
      "calib/mean_conf": 0.04792156862745098,
      "calib/mu_c": 0.05056451612903227,
      "calib/mu_w": 0.04541984732824427,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.014224865190794278,
      "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": 805.0,
      "completions/max_terminated_length": 805.0,
      "completions/mean_length": 313.69140625,
      "completions/mean_terminated_length": 314.9216003417969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 125.0,
      "epoch": 0.15146666666666667,
      "grad_norm": 0.009747452102601528,
      "learning_rate": 1.6111111111111113e-06,
      "loss": 0.0442,
      "num_tokens": 25647996.0,
      "reward": 1.0172728300094604,
      "reward_std": 0.03187453746795654,
      "rewards/accuracy_reward_step": 0.484375,
      "rewards/final_brier_reward_step": 0.5579832196235657,
      "rewards/format_reward_step": 0.9921875,
      "step": 142
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3411660713609308,
      "aux_distill/mean_u": 0.1608952434332203,
      "aux_distill/n_active_tok": 262.4375,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5976733977172958,
      "calib/avg_num_step_conf": 4.09765625,
      "calib/ece": 0.4187747035573122,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00510660980810234,
      "calib/mean_conf": 0.05158102766798419,
      "calib/mu_c": 0.05428571428571428,
      "calib/mu_w": 0.04917910447761194,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.0,
      "calib/std_conf": 0.017999515686590115,
      "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": 2703.0,
      "completions/max_terminated_length": 2703.0,
      "completions/mean_length": 365.1171875,
      "completions/mean_terminated_length": 365.1171875,
      "completions/min_length": 101.0,
      "completions/min_terminated_length": 101.0,
      "epoch": 0.15253333333333333,
      "grad_norm": 0.008262704126536846,
      "learning_rate": 1.5833333333333333e-06,
      "loss": 0.0777,
      "num_tokens": 25848802.0,
      "reward": 1.0081377029418945,
      "reward_std": 0.05720390006899834,
      "rewards/accuracy_reward_step": 0.46484375,
      "rewards/final_brier_reward_step": 0.5670566558837891,
      "rewards/format_reward_step": 0.984375,
      "step": 143
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3373702000826597,
      "aux_distill/mean_u": 0.13485213419125378,
      "aux_distill/n_active_tok": 263.46875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5864911417322834,
      "calib/avg_num_step_conf": 4.109375,
      "calib/ece": 0.44854901960784316,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.007110605314960626,
      "calib/mean_conf": 0.05341176470588236,
      "calib/mu_c": 0.05695312500000001,
      "calib/mu_w": 0.04984251968503938,
      "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": 0.021358909121174646,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2474.0,
      "completions/max_terminated_length": 2474.0,
      "completions/mean_length": 334.23046875,
      "completions/mean_terminated_length": 334.23046875,
      "completions/min_length": 99.0,
      "completions/min_terminated_length": 99.0,
      "epoch": 0.1536,
      "grad_norm": 0.00993384700268507,
      "learning_rate": 1.5555555555555558e-06,
      "loss": 0.0846,
      "num_tokens": 26038493.0,
      "reward": 1.0229222774505615,
      "reward_std": 0.03093772381544113,
      "rewards/accuracy_reward_step": 0.5,
      "rewards/final_brier_reward_step": 0.5497507452964783,
      "rewards/format_reward_step": 0.99609375,
      "step": 144
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34878645557910204,
      "aux_distill/mean_u": 0.18183138162863954,
      "aux_distill/n_active_tok": 289.125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5883374689826303,
      "calib/avg_num_step_conf": 4.5078125,
      "calib/ece": 0.45669291338582674,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.006533498759305208,
      "calib/mean_conf": 0.05511811023622047,
      "calib/mu_c": 0.05830769230769231,
      "calib/mu_w": 0.0517741935483871,
      "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": 0.01774119019399936,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2132.0,
      "completions/max_terminated_length": 2132.0,
      "completions/mean_length": 311.1015625,
      "completions/mean_terminated_length": 312.32159423828125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 87.0,
      "epoch": 0.15466666666666667,
      "grad_norm": 0.009975967928767204,
      "learning_rate": 1.527777777777778e-06,
      "loss": 0.0247,
      "num_tokens": 26220839.0,
      "reward": 1.0201336145401,
      "reward_std": 0.036704327911138535,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.5402672290802002,
      "rewards/format_reward_step": 0.9921875,
      "step": 145
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.398836228530854,
      "aux_distill/mean_u": 0.17221634645466294,
      "aux_distill/n_active_tok": 287.5625,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5221437740693197,
      "calib/avg_num_step_conf": 4.4921875,
      "calib/ece": 0.27276679841897233,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0004671230922835487,
      "calib/mean_conf": 0.05395256916996048,
      "calib/mu_c": 0.05426829268292682,
      "calib/mu_w": 0.05380116959064327,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.0013043478260869562,
      "calib/std_conf": 0.017563822730073224,
      "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": 2142.0,
      "completions/max_terminated_length": 2142.0,
      "completions/mean_length": 347.65625,
      "completions/mean_terminated_length": 349.0196228027344,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 65.0,
      "epoch": 0.15573333333333333,
      "grad_norm": 0.009122155606746674,
      "learning_rate": 1.5e-06,
      "loss": 0.0681,
      "num_tokens": 26417055.0,
      "reward": 1.0001678466796875,
      "reward_std": 0.05830451473593712,
      "rewards/accuracy_reward_step": 0.3203125,
      "rewards/final_brier_reward_step": 0.6956480741500854,
      "rewards/format_reward_step": 0.984375,
      "step": 146
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33496221201494336,
      "aux_distill/mean_u": 0.13949762871699134,
      "aux_distill/n_active_tok": 265.6875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5589768654284784,
      "calib/avg_num_step_conf": 4.20703125,
      "calib/ece": 0.33366141732283466,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0030765721733463716,
      "calib/mean_conf": 0.05610236220472441,
      "calib/mu_c": 0.05797979797979799,
      "calib/mu_w": 0.05490322580645162,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.016629513013958232,
      "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": 2389.0,
      "completions/max_terminated_length": 2389.0,
      "completions/mean_length": 319.578125,
      "completions/mean_terminated_length": 320.8313903808594,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 97.0,
      "epoch": 0.1568,
      "grad_norm": 0.010258171707391739,
      "learning_rate": 1.4722222222222225e-06,
      "loss": 0.0565,
      "num_tokens": 26602547.0,
      "reward": 1.0108044147491455,
      "reward_std": 0.04217313602566719,
      "rewards/accuracy_reward_step": 0.38671875,
      "rewards/final_brier_reward_step": 0.6466090083122253,
      "rewards/format_reward_step": 0.98828125,
      "step": 147
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34962216787971556,
      "aux_distill/mean_u": 0.18060808775131193,
      "aux_distill/n_active_tok": 267.46875,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6791163382072471,
      "calib/avg_num_step_conf": 4.16796875,
      "calib/ece": 0.5073517786561265,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.011825174825174818,
      "calib/mean_conf": 0.05786561264822135,
      "calib/mu_c": 0.063006993006993,
      "calib/mu_w": 0.05118181818181818,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.020874879548542648,
      "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": 2560.0,
      "completions/max_terminated_length": 2560.0,
      "completions/mean_length": 335.19140625,
      "completions/mean_terminated_length": 335.19140625,
      "completions/min_length": 88.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.15786666666666666,
      "grad_norm": 0.01002125721424818,
      "learning_rate": 1.4444444444444445e-06,
      "loss": 0.119,
      "num_tokens": 26793468.0,
      "reward": 1.0118077993392944,
      "reward_std": 0.07325895875692368,
      "rewards/accuracy_reward_step": 0.55859375,
      "rewards/final_brier_reward_step": 0.48845937848091125,
      "rewards/format_reward_step": 0.9765625,
      "step": 148
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3421163931488991,
      "aux_distill/mean_u": 0.15572480646023862,
      "aux_distill/n_active_tok": 286.1875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6150439737396259,
      "calib/avg_num_step_conf": 4.48828125,
      "calib/ece": 0.40062745098039215,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.010598290598290608,
      "calib/mean_conf": 0.05819607843137255,
      "calib/mu_c": 0.06393162393162394,
      "calib/mu_w": 0.05333333333333333,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.022610963511000795,
      "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": 932.0,
      "completions/max_terminated_length": 932.0,
      "completions/mean_length": 352.83203125,
      "completions/mean_terminated_length": 354.2156982421875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 76.0,
      "epoch": 0.15893333333333334,
      "grad_norm": 0.00943607185035944,
      "learning_rate": 1.4166666666666667e-06,
      "loss": 0.059,
      "num_tokens": 26988249.0,
      "reward": 1.015564203262329,
      "reward_std": 0.048596229404211044,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/final_brier_reward_step": 0.5858160257339478,
      "rewards/format_reward_step": 0.98828125,
      "step": 149
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36728407302871346,
      "aux_distill/mean_u": 0.17260016591145952,
      "aux_distill/n_active_tok": 240.28125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6337993319312137,
      "calib/avg_num_step_conf": 3.73828125,
      "calib/ece": 0.39545098039215687,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.012134727205245568,
      "calib/mean_conf": 0.06729411764705882,
      "calib/mu_c": 0.0738135593220339,
      "calib/mu_w": 0.06167883211678833,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.027427833378327503,
      "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": 1790.0,
      "completions/max_terminated_length": 1790.0,
      "completions/mean_length": 281.06640625,
      "completions/mean_terminated_length": 281.06640625,
      "completions/min_length": 104.0,
      "completions/min_terminated_length": 104.0,
      "epoch": 0.16,
      "grad_norm": 0.011365544982254505,
      "learning_rate": 1.3888888888888892e-06,
      "loss": 0.0586,
      "num_tokens": 27165162.0,
      "reward": 1.02748703956604,
      "reward_std": 0.0358702652156353,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/final_brier_reward_step": 0.5979429483413696,
      "rewards/format_reward_step": 0.99609375,
      "step": 150
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37775949435308576,
      "aux_distill/mean_u": 0.19691793188973097,
      "aux_distill/n_active_tok": 308.34375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5913930789707187,
      "calib/avg_num_step_conf": 4.8125,
      "calib/ece": 0.2962301587301587,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.007668418537983772,
      "calib/mean_conf": 0.06488095238095239,
      "calib/mu_c": 0.0697802197802198,
      "calib/mu_w": 0.062111801242236024,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.98828125,
      "calib/pce": 0.0,
      "calib/std_conf": 0.025811928810808394,
      "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": 2308.0,
      "completions/max_terminated_length": 2308.0,
      "completions/mean_length": 372.53515625,
      "completions/mean_terminated_length": 372.53515625,
      "completions/min_length": 117.0,
      "completions/min_terminated_length": 117.0,
      "epoch": 0.16106666666666666,
      "grad_norm": 0.008531233295798302,
      "learning_rate": 1.3611111111111112e-06,
      "loss": 0.0956,
      "num_tokens": 27367555.0,
      "reward": 1.0047494173049927,
      "reward_std": 0.07198721170425415,
      "rewards/accuracy_reward_step": 0.35546875,
      "rewards/final_brier_reward_step": 0.6735613346099854,
      "rewards/format_reward_step": 0.98046875,
      "step": 151
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3672543163411319,
      "aux_distill/mean_u": 0.13673712651834333,
      "aux_distill/n_active_tok": 297.3125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5581105870020965,
      "calib/avg_num_step_conf": 4.64453125,
      "calib/ece": 0.3008235294117647,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.00392156862745098,
      "calib/gap": 0.01533411949685537,
      "calib/mean_conf": 0.07564705882352941,
      "calib/mu_c": 0.08520833333333334,
      "calib/mu_w": 0.06987421383647797,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.06668827792736973,
      "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": 1254.0,
      "completions/max_terminated_length": 1254.0,
      "completions/mean_length": 312.6640625,
      "completions/mean_terminated_length": 313.8902282714844,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 99.0,
      "epoch": 0.16213333333333332,
      "grad_norm": 0.009770411998033524,
      "learning_rate": 1.3333333333333334e-06,
      "loss": -0.0123,
      "num_tokens": 27552989.0,
      "reward": 1.019080400466919,
      "reward_std": 0.049445126205682755,
      "rewards/accuracy_reward_step": 0.375,
      "rewards/final_brier_reward_step": 0.6709734201431274,
      "rewards/format_reward_step": 0.9921875,
      "step": 152
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38715670723468065,
      "aux_distill/mean_u": 0.2227721838341249,
      "aux_distill/n_active_tok": 298.53125,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.6068790819752183,
      "calib/avg_num_step_conf": 4.6015625,
      "calib/ece": 0.4351953125,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.010723921137764755,
      "calib/mean_conf": 0.06871093750000001,
      "calib/mu_c": 0.074031007751938,
      "calib/mu_w": 0.06330708661417324,
      "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": 0.029795829286514138,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1053.0,
      "completions/max_terminated_length": 1053.0,
      "completions/mean_length": 323.4921875,
      "completions/mean_terminated_length": 324.76080322265625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 123.0,
      "epoch": 0.1632,
      "grad_norm": 0.010421986691653728,
      "learning_rate": 1.3055555555555556e-06,
      "loss": 0.0204,
      "num_tokens": 27743123.0,
      "reward": 1.034500241279602,
      "reward_std": 0.027013670653104782,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.5650941133499146,
      "rewards/format_reward_step": 1.0,
      "step": 153
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35084733460098505,
      "aux_distill/mean_u": 0.16063960511918457,
      "aux_distill/n_active_tok": 298.09375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5794099378881987,
      "calib/avg_num_step_conf": 4.65234375,
      "calib/ece": 0.3744705882352941,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.010627329192546595,
      "calib/mean_conf": 0.07807843137254902,
      "calib/mu_c": 0.08391304347826087,
      "calib/mu_w": 0.07328571428571427,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.0007843137254901962,
      "calib/std_conf": 0.040494900737928746,
      "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": 766.0,
      "completions/max_terminated_length": 766.0,
      "completions/mean_length": 303.015625,
      "completions/mean_terminated_length": 304.2039489746094,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 111.0,
      "epoch": 0.16426666666666667,
      "grad_norm": 0.010418664664030075,
      "learning_rate": 1.2777777777777779e-06,
      "loss": 0.0423,
      "num_tokens": 27925135.0,
      "reward": 1.0232136249542236,
      "reward_std": 0.051645420491695404,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.6128336191177368,
      "rewards/format_reward_step": 0.984375,
      "step": 154
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33021269901655614,
      "aux_distill/mean_u": 0.14969116715629144,
      "aux_distill/n_active_tok": 292.75,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5542440660474716,
      "calib/avg_num_step_conf": 4.54296875,
      "calib/ece": 0.32307086614173225,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.005285087719298254,
      "calib/mean_conf": 0.07850393700787402,
      "calib/mu_c": 0.08166666666666667,
      "calib/mu_w": 0.07638157894736841,
      "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": 0.03686704650983364,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2633.0,
      "completions/max_terminated_length": 2633.0,
      "completions/mean_length": 304.2734375,
      "completions/mean_terminated_length": 304.2734375,
      "completions/min_length": 91.0,
      "completions/min_terminated_length": 91.0,
      "epoch": 0.16533333333333333,
      "grad_norm": 0.011450732126832008,
      "learning_rate": 1.25e-06,
      "loss": 0.0748,
      "num_tokens": 28110245.0,
      "reward": 1.0229480266571045,
      "reward_std": 0.04218357801437378,
      "rewards/accuracy_reward_step": 0.40234375,
      "rewards/final_brier_reward_step": 0.6513648629188538,
      "rewards/format_reward_step": 0.9921875,
      "step": 155
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3649008311331272,
      "aux_distill/mean_u": 0.16765184443474762,
      "aux_distill/n_active_tok": 335.375,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5861605428242223,
      "calib/avg_num_step_conf": 5.22265625,
      "calib/ece": 0.34644268774703557,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.003197413903469476,
      "calib/mean_conf": 0.08320158102766798,
      "calib/mu_c": 0.08504672897196262,
      "calib/mu_w": 0.08184931506849315,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0033596837944664033,
      "calib/std_conf": 0.04752809526372359,
      "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": 2478.0,
      "completions/max_terminated_length": 2478.0,
      "completions/mean_length": 340.5703125,
      "completions/mean_terminated_length": 340.5703125,
      "completions/min_length": 110.0,
      "completions/min_terminated_length": 110.0,
      "epoch": 0.1664,
      "grad_norm": 0.008358205668628216,
      "learning_rate": 1.2222222222222223e-06,
      "loss": 0.067,
      "num_tokens": 28302191.0,
      "reward": 1.0192911624908447,
      "reward_std": 0.05631423741579056,
      "rewards/accuracy_reward_step": 0.41796875,
      "rewards/final_brier_reward_step": 0.632332444190979,
      "rewards/format_reward_step": 0.98828125,
      "step": 156
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38443474285304546,
      "aux_distill/mean_u": 0.19694286921091003,
      "aux_distill/n_active_tok": 322.59375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6189921259842519,
      "calib/avg_num_step_conf": 5.01953125,
      "calib/ece": 0.4149603174603174,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.01835023622047245,
      "calib/mean_conf": 0.08900793650793651,
      "calib/mu_c": 0.09811023622047248,
      "calib/mu_w": 0.07976000000000003,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.984375,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.04890272106425196,
      "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": 3031.0,
      "completions/max_terminated_length": 3031.0,
      "completions/mean_length": 335.7890625,
      "completions/mean_terminated_length": 337.10589599609375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 76.0,
      "epoch": 0.16746666666666668,
      "grad_norm": 0.008726774714887142,
      "learning_rate": 1.1944444444444446e-06,
      "loss": 0.1026,
      "num_tokens": 28491881.0,
      "reward": 1.027970552444458,
      "reward_std": 0.0656554326415062,
      "rewards/accuracy_reward_step": 0.49609375,
      "rewards/final_brier_reward_step": 0.5754722356796265,
      "rewards/format_reward_step": 0.984375,
      "step": 157
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3361390228383243,
      "aux_distill/mean_u": 0.1681130252148755,
      "aux_distill/n_active_tok": 342.53125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.6021364317841078,
      "calib/avg_num_step_conf": 5.33203125,
      "calib/ece": 0.4535826771653544,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.012662418790604688,
      "calib/mean_conf": 0.0897244094488189,
      "calib/mu_c": 0.09550724637681159,
      "calib/mu_w": 0.0828448275862069,
      "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": 0.03995473367151239,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1953.0,
      "completions/max_terminated_length": 1953.0,
      "completions/mean_length": 334.80078125,
      "completions/mean_terminated_length": 334.80078125,
      "completions/min_length": 83.0,
      "completions/min_terminated_length": 83.0,
      "epoch": 0.16853333333333334,
      "grad_norm": 0.009793547913432121,
      "learning_rate": 1.1666666666666668e-06,
      "loss": 0.0737,
      "num_tokens": 28682830.0,
      "reward": 1.0388861894607544,
      "reward_std": 0.05535713583230972,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/final_brier_reward_step": 0.5465222597122192,
      "rewards/format_reward_step": 0.9921875,
      "step": 158
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34734752285294235,
      "aux_distill/mean_u": 0.14985655571886566,
      "aux_distill/n_active_tok": 300.0,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5640445314411549,
      "calib/avg_num_step_conf": 4.73828125,
      "calib/ece": 0.383515625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.005874724736970871,
      "calib/mean_conf": 0.09585937500000002,
      "calib/mu_c": 0.0989344262295082,
      "calib/mu_w": 0.09305970149253733,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.00140625,
      "calib/std_conf": 0.042912107552640376,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1069.0,
      "completions/max_terminated_length": 1069.0,
      "completions/mean_length": 288.9375,
      "completions/mean_terminated_length": 290.07061767578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 89.0,
      "epoch": 0.1696,
      "grad_norm": 0.009990203194320202,
      "learning_rate": 1.138888888888889e-06,
      "loss": 0.0227,
      "num_tokens": 28861582.0,
      "reward": 1.0416332483291626,
      "reward_std": 0.03353843465447426,
      "rewards/accuracy_reward_step": 0.4765625,
      "rewards/final_brier_reward_step": 0.6067039370536804,
      "rewards/format_reward_step": 1.0,
      "step": 159
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36702860286459327,
      "aux_distill/mean_u": 0.16000241486172087,
      "aux_distill/n_active_tok": 336.375,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5723726627981947,
      "calib/avg_num_step_conf": 5.1875,
      "calib/ece": 0.3390836653386454,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.009315925209542247,
      "calib/mean_conf": 0.10649402390438246,
      "calib/mu_c": 0.11172727272727276,
      "calib/mu_w": 0.10241134751773051,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.003665338645418327,
      "calib/std_conf": 0.0613433528297715,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2485.0,
      "completions/max_terminated_length": 2485.0,
      "completions/mean_length": 345.75,
      "completions/mean_terminated_length": 345.75,
      "completions/min_length": 105.0,
      "completions/min_terminated_length": 105.0,
      "epoch": 0.17066666666666666,
      "grad_norm": 0.00989892240613699,
      "learning_rate": 1.111111111111111e-06,
      "loss": 0.0859,
      "num_tokens": 29054934.0,
      "reward": 1.0210720300674438,
      "reward_std": 0.0724157989025116,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.6319879293441772,
      "rewards/format_reward_step": 0.98046875,
      "step": 160
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33740366343408823,
      "aux_distill/mean_u": 0.17007575680141462,
      "aux_distill/n_active_tok": 324.375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5586751626178148,
      "calib/avg_num_step_conf": 5.078125,
      "calib/ece": 0.5285490196078432,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.009093986459577874,
      "calib/mean_conf": 0.10674509803921568,
      "calib/mu_c": 0.11006172839506174,
      "calib/mu_w": 0.10096774193548387,
      "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": 0.0516776749071968,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1226.0,
      "completions/max_terminated_length": 1226.0,
      "completions/mean_length": 315.41015625,
      "completions/mean_terminated_length": 316.6470642089844,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 45.0,
      "epoch": 0.17173333333333332,
      "grad_norm": 0.009658781811594963,
      "learning_rate": 1.0833333333333335e-06,
      "loss": 0.0206,
      "num_tokens": 29239599.0,
      "reward": 1.05873703956604,
      "reward_std": 0.05385105684399605,
      "rewards/accuracy_reward_step": 0.6328125,
      "rewards/final_brier_reward_step": 0.48856794834136963,
      "rewards/format_reward_step": 0.99609375,
      "step": 161
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.351145098451525,
      "aux_distill/mean_u": 0.16597964091285583,
      "aux_distill/n_active_tok": 344.96875,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.6000405844155845,
      "calib/avg_num_step_conf": 5.375,
      "calib/ece": 0.5466015625000001,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.016260822510822517,
      "calib/mean_conf": 0.10964843750000003,
      "calib/mu_c": 0.11523809523809525,
      "calib/mu_w": 0.09897727272727273,
      "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": 0.054988221728008206,
      "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": 1110.0,
      "completions/max_terminated_length": 1110.0,
      "completions/mean_length": 315.31640625,
      "completions/mean_terminated_length": 316.5529479980469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 100.0,
      "epoch": 0.1728,
      "grad_norm": 0.009903507307171822,
      "learning_rate": 1.0555555555555557e-06,
      "loss": 0.0578,
      "num_tokens": 29424464.0,
      "reward": 1.0642025470733643,
      "reward_std": 0.05435194447636604,
      "rewards/accuracy_reward_step": 0.65625,
      "rewards/final_brier_reward_step": 0.47606131434440613,
      "rewards/format_reward_step": 0.99609375,
      "step": 162
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3831592188216746,
      "aux_distill/mean_u": 0.19678231264481222,
      "aux_distill/n_active_tok": 386.3125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.6116102499209112,
      "calib/avg_num_step_conf": 6.0078125,
      "calib/ece": 0.31413385826771656,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.021293894337235053,
      "calib/mean_conf": 0.115,
      "calib/mu_c": 0.1271559633027523,
      "calib/mu_w": 0.10586206896551725,
      "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": 0.0638511433252139,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2184.0,
      "completions/max_terminated_length": 2184.0,
      "completions/mean_length": 355.8203125,
      "completions/mean_terminated_length": 355.8203125,
      "completions/min_length": 88.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.17386666666666667,
      "grad_norm": 0.009598040953278542,
      "learning_rate": 1.0277777777777777e-06,
      "loss": 0.0468,
      "num_tokens": 29620386.0,
      "reward": 1.0377447605133057,
      "reward_std": 0.06843307614326477,
      "rewards/accuracy_reward_step": 0.42578125,
      "rewards/final_brier_reward_step": 0.6575207114219666,
      "rewards/format_reward_step": 0.9921875,
      "step": 163
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.361350919585675,
      "aux_distill/mean_u": 0.1579362222322398,
      "aux_distill/n_active_tok": 359.28125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6311188811188811,
      "calib/avg_num_step_conf": 5.59765625,
      "calib/ece": 0.2724313725490196,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.026507381507381494,
      "calib/mean_conf": 0.11580392156862745,
      "calib/mu_c": 0.132020202020202,
      "calib/mu_w": 0.10551282051282052,
      "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": 0.05963319894147763,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1842.0,
      "completions/max_terminated_length": 1842.0,
      "completions/mean_length": 346.69921875,
      "completions/mean_terminated_length": 346.69921875,
      "completions/min_length": 97.0,
      "completions/min_terminated_length": 97.0,
      "epoch": 0.17493333333333333,
      "grad_norm": 0.010199583135545254,
      "learning_rate": 1.0000000000000002e-06,
      "loss": 0.0409,
      "num_tokens": 29815277.0,
      "reward": 1.0386981964111328,
      "reward_std": 0.059402212500572205,
      "rewards/accuracy_reward_step": 0.38671875,
      "rewards/final_brier_reward_step": 0.6945840120315552,
      "rewards/format_reward_step": 0.99609375,
      "step": 164
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3957953127101064,
      "aux_distill/mean_u": 0.17184055225291792,
      "aux_distill/n_active_tok": 346.4375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5459833614974652,
      "calib/avg_num_step_conf": 5.37890625,
      "calib/ece": 0.2625490196078431,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.007409333160015602,
      "calib/mean_conf": 0.12176470588235294,
      "calib/mu_c": 0.1263265306122449,
      "calib/mu_w": 0.11891719745222931,
      "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": 0.06361359708991009,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 3020.0,
      "completions/max_terminated_length": 3020.0,
      "completions/mean_length": 347.6171875,
      "completions/mean_terminated_length": 347.6171875,
      "completions/min_length": 94.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.176,
      "grad_norm": 0.010400495491921902,
      "learning_rate": 9.722222222222224e-07,
      "loss": 0.0696,
      "num_tokens": 30009843.0,
      "reward": 1.0350532531738281,
      "reward_std": 0.05347467213869095,
      "rewards/accuracy_reward_step": 0.3828125,
      "rewards/final_brier_reward_step": 0.6912003755569458,
      "rewards/format_reward_step": 0.99609375,
      "step": 165
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.37643085792660713,
      "aux_distill/mean_u": 0.2064087180197083,
      "aux_distill/n_active_tok": 397.71875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.6964863184079603,
      "calib/avg_num_step_conf": 6.15625,
      "calib/ece": 0.40870078740157484,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.03852238805970147,
      "calib/mean_conf": 0.1223228346456693,
      "calib/mu_c": 0.1405223880597015,
      "calib/mu_w": 0.10200000000000002,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.001732283464566929,
      "calib/std_conf": 0.06698012184725738,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2423.0,
      "completions/max_terminated_length": 2423.0,
      "completions/mean_length": 390.49609375,
      "completions/mean_terminated_length": 390.49609375,
      "completions/min_length": 108.0,
      "completions/min_terminated_length": 108.0,
      "epoch": 0.17706666666666668,
      "grad_norm": 0.00810649711638689,
      "learning_rate": 9.444444444444445e-07,
      "loss": 0.0979,
      "num_tokens": 30215994.0,
      "reward": 1.0560935735702515,
      "reward_std": 0.07129926979541779,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/final_brier_reward_step": 0.5965621471405029,
      "rewards/format_reward_step": 0.9921875,
      "step": 166
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38002316700294614,
      "aux_distill/mean_u": 0.19313883956834368,
      "aux_distill/n_active_tok": 386.53125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6171767214348426,
      "calib/avg_num_step_conf": 6.01953125,
      "calib/ece": 0.330197628458498,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.033337514110121605,
      "calib/mean_conf": 0.1401581027667984,
      "calib/mu_c": 0.1578151260504201,
      "calib/mu_w": 0.12447761194029851,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.08695993525459589,
      "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": 2277.0,
      "completions/max_terminated_length": 2277.0,
      "completions/mean_length": 365.75390625,
      "completions/mean_terminated_length": 365.75390625,
      "completions/min_length": 113.0,
      "completions/min_terminated_length": 113.0,
      "epoch": 0.17813333333333334,
      "grad_norm": 0.008568638935685158,
      "learning_rate": 9.166666666666666e-07,
      "loss": 0.0978,
      "num_tokens": 30415235.0,
      "reward": 1.0457689762115479,
      "reward_std": 0.09472964704036713,
      "rewards/accuracy_reward_step": 0.46484375,
      "rewards/final_brier_reward_step": 0.6423191428184509,
      "rewards/format_reward_step": 0.984375,
      "step": 167
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3776658959686756,
      "aux_distill/mean_u": 0.21580393844564622,
      "aux_distill/n_active_tok": 337.90625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.6339692307692307,
      "calib/avg_num_step_conf": 5.27734375,
      "calib/ece": 0.3627843137254902,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.03651076923076921,
      "calib/mean_conf": 0.14733333333333334,
      "calib/mu_c": 0.16523076923076924,
      "calib/mu_w": 0.12872000000000003,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.00015686274509803936,
      "calib/std_conf": 0.0849910414348654,
      "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": 932.0,
      "completions/max_terminated_length": 932.0,
      "completions/mean_length": 335.875,
      "completions/mean_terminated_length": 337.1921691894531,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 66.0,
      "epoch": 0.1792,
      "grad_norm": 0.009574908763170242,
      "learning_rate": 8.88888888888889e-07,
      "loss": 0.0351,
      "num_tokens": 30605891.0,
      "reward": 1.0578217506408691,
      "reward_std": 0.08997109532356262,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.6195499897003174,
      "rewards/format_reward_step": 0.98828125,
      "step": 168
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3557533877901733,
      "aux_distill/mean_u": 0.15889651282363162,
      "aux_distill/n_active_tok": 359.59375,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6078159645232816,
      "calib/avg_num_step_conf": 5.59765625,
      "calib/ece": 0.3290980392156862,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.025280857354028102,
      "calib/mean_conf": 0.1532549019607843,
      "calib/mu_c": 0.16634146341463416,
      "calib/mu_w": 0.14106060606060605,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.08227163119508565,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2253.0,
      "completions/max_terminated_length": 2253.0,
      "completions/mean_length": 355.71875,
      "completions/mean_terminated_length": 355.71875,
      "completions/min_length": 71.0,
      "completions/min_terminated_length": 71.0,
      "epoch": 0.18026666666666666,
      "grad_norm": 0.009443276561796665,
      "learning_rate": 8.611111111111112e-07,
      "loss": 0.04,
      "num_tokens": 30801139.0,
      "reward": 1.051137089729309,
      "reward_std": 0.0901157557964325,
      "rewards/accuracy_reward_step": 0.48046875,
      "rewards/final_brier_reward_step": 0.6374304294586182,
      "rewards/format_reward_step": 0.984375,
      "step": 169
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.351165306288749,
      "aux_distill/mean_u": 0.1616900184663727,
      "aux_distill/n_active_tok": 420.21875,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.5186444948349711,
      "calib/avg_num_step_conf": 6.58203125,
      "calib/ece": 0.33551587301587305,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.005952380952380959,
      "calib/mean_conf": 0.16829365079365083,
      "calib/mu_c": 0.17126984126984127,
      "calib/mu_w": 0.1653174603174603,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.001904761904761905,
      "calib/std_conf": 0.08327536721674905,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2416.0,
      "completions/max_terminated_length": 2416.0,
      "completions/mean_length": 377.94921875,
      "completions/mean_terminated_length": 379.431396484375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 113.0,
      "epoch": 0.18133333333333335,
      "grad_norm": 0.019583020359277725,
      "learning_rate": 8.333333333333333e-07,
      "loss": 0.067,
      "num_tokens": 31002046.0,
      "reward": 1.051318645477295,
      "reward_std": 0.09664470702409744,
      "rewards/accuracy_reward_step": 0.4921875,
      "rewards/final_brier_reward_step": 0.6260746121406555,
      "rewards/format_reward_step": 0.984375,
      "step": 170
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3422384618315846,
      "aux_distill/mean_u": 0.1775502851850448,
      "aux_distill/n_active_tok": 346.15625,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5632081280788178,
      "calib/avg_num_step_conf": 5.40234375,
      "calib/ece": 0.28039062499999995,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.026197044334975356,
      "calib/mean_conf": 0.17765625000000002,
      "calib/mu_c": 0.19198275862068964,
      "calib/mu_w": 0.1657857142857143,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0024609375,
      "calib/std_conf": 0.09883686349706522,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 928.0,
      "completions/max_terminated_length": 928.0,
      "completions/mean_length": 334.4609375,
      "completions/mean_terminated_length": 335.7725830078125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 98.0,
      "epoch": 0.1824,
      "grad_norm": 0.009695254266262054,
      "learning_rate": 8.055555555555557e-07,
      "loss": 0.0486,
      "num_tokens": 31194564.0,
      "reward": 1.0663268566131592,
      "reward_std": 0.08017165958881378,
      "rewards/accuracy_reward_step": 0.453125,
      "rewards/final_brier_reward_step": 0.6795288920402527,
      "rewards/format_reward_step": 1.0,
      "step": 171
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34708446078002453,
      "aux_distill/mean_u": 0.12701413012860135,
      "aux_distill/n_active_tok": 359.28125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5189120139581256,
      "calib/avg_num_step_conf": 5.57421875,
      "calib/ece": 0.34846456692913386,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.007395314057826563,
      "calib/mean_conf": 0.19641732283464566,
      "calib/mu_c": 0.1998529411764706,
      "calib/mu_w": 0.19245762711864403,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.004724409448818897,
      "calib/std_conf": 0.09339273951655766,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1859.0,
      "completions/max_terminated_length": 1859.0,
      "completions/mean_length": 331.703125,
      "completions/mean_terminated_length": 331.703125,
      "completions/min_length": 103.0,
      "completions/min_terminated_length": 103.0,
      "epoch": 0.18346666666666667,
      "grad_norm": 0.009064041078090668,
      "learning_rate": 7.777777777777779e-07,
      "loss": 0.0595,
      "num_tokens": 31382832.0,
      "reward": 1.0748931169509888,
      "reward_std": 0.09136275202035904,
      "rewards/accuracy_reward_step": 0.53125,
      "rewards/final_brier_reward_step": 0.6263488531112671,
      "rewards/format_reward_step": 0.9921875,
      "step": 172
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.31844893423840404,
      "aux_distill/mean_u": 0.1626621419976694,
      "aux_distill/n_active_tok": 407.125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.4622554448135844,
      "calib/avg_num_step_conf": 6.32421875,
      "calib/ece": 0.33415686274509815,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": -0.019946474713916584,
      "calib/mean_conf": 0.19094117647058823,
      "calib/mu_c": 0.18108527131782945,
      "calib/mu_w": 0.20103174603174603,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.009607843137254903,
      "calib/std_conf": 0.10646870326573207,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2345.0,
      "completions/max_terminated_length": 2345.0,
      "completions/mean_length": 368.19140625,
      "completions/mean_terminated_length": 368.19140625,
      "completions/min_length": 72.0,
      "completions/min_terminated_length": 72.0,
      "epoch": 0.18453333333333333,
      "grad_norm": 0.008606944233179092,
      "learning_rate": 7.5e-07,
      "loss": 0.0435,
      "num_tokens": 31580249.0,
      "reward": 1.0635401010513306,
      "reward_std": 0.08568964153528214,
      "rewards/accuracy_reward_step": 0.50390625,
      "rewards/final_brier_reward_step": 0.6270800828933716,
      "rewards/format_reward_step": 0.99609375,
      "step": 173
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.379526324570179,
      "aux_distill/mean_u": 0.19726045253259963,
      "aux_distill/n_active_tok": 413.875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.47033685268979386,
      "calib/avg_num_step_conf": 6.45703125,
      "calib/ece": 0.2892094861660079,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.003952569169960474,
      "calib/gap": -0.015221216691804956,
      "calib/mean_conf": 0.1850197628458498,
      "calib/mu_c": 0.17683760683760683,
      "calib/mu_w": 0.19205882352941178,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.005889328063241106,
      "calib/std_conf": 0.10748885227953005,
      "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": 2196.0,
      "completions/max_terminated_length": 2196.0,
      "completions/mean_length": 396.53515625,
      "completions/mean_terminated_length": 396.53515625,
      "completions/min_length": 94.0,
      "completions/min_terminated_length": 94.0,
      "epoch": 0.1856,
      "grad_norm": 0.008462514728307724,
      "learning_rate": 7.222222222222222e-07,
      "loss": 0.087,
      "num_tokens": 31785994.0,
      "reward": 1.0464767217636108,
      "reward_std": 0.10494433343410492,
      "rewards/accuracy_reward_step": 0.45703125,
      "rewards/final_brier_reward_step": 0.6476410627365112,
      "rewards/format_reward_step": 0.98828125,
      "step": 174
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3483807328157127,
      "aux_distill/mean_u": 0.16240182766074354,
      "aux_distill/n_active_tok": 386.625,
      "calib/answer_extract_rate": 0.98046875,
      "calib/auroc": 0.5680039138943248,
      "calib/avg_num_step_conf": 6.01953125,
      "calib/ece": 0.2311553784860558,
      "calib/final_conf_rate": 0.98046875,
      "calib/format_rate": 0.96484375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.024359425962165643,
      "calib/mean_conf": 0.2013545816733068,
      "calib/mu_c": 0.21552380952380948,
      "calib/mu_w": 0.19116438356164384,
      "calib/nonempty_final_conf_rate": 0.98046875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.007091633466135458,
      "calib/std_conf": 0.11309023020690735,
      "calib/step_conf_rate": 0.984375,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.0,
      "completions/max_length": 2076.0,
      "completions/max_terminated_length": 2076.0,
      "completions/mean_length": 348.3125,
      "completions/mean_terminated_length": 348.3125,
      "completions/min_length": 88.0,
      "completions/min_terminated_length": 88.0,
      "epoch": 0.18666666666666668,
      "grad_norm": 0.009568187408149242,
      "learning_rate": 6.944444444444446e-07,
      "loss": 0.02,
      "num_tokens": 31980986.0,
      "reward": 1.0292150974273682,
      "reward_std": 0.14621660113334656,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.6834300756454468,
      "rewards/format_reward_step": 0.96484375,
      "step": 175
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.390253531280905,
      "aux_distill/mean_u": 0.23588810433493573,
      "aux_distill/n_active_tok": 391.28125,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.609801031687546,
      "calib/avg_num_step_conf": 6.1484375,
      "calib/ece": 0.24800781249999998,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.04109923851633504,
      "calib/mean_conf": 0.21292968750000002,
      "calib/mu_c": 0.23508474576271188,
      "calib/mu_w": 0.19398550724637684,
      "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": 0.10612313629059567,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1037.0,
      "completions/max_terminated_length": 1037.0,
      "completions/mean_length": 349.96875,
      "completions/mean_terminated_length": 351.3411865234375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 106.0,
      "epoch": 0.18773333333333334,
      "grad_norm": 0.010521200485527515,
      "learning_rate": 6.666666666666667e-07,
      "loss": 0.0337,
      "num_tokens": 32174642.0,
      "reward": 1.0800588130950928,
      "reward_std": 0.07746738940477371,
      "rewards/accuracy_reward_step": 0.4609375,
      "rewards/final_brier_reward_step": 0.6991801261901855,
      "rewards/format_reward_step": 1.0,
      "step": 176
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34870817279443145,
      "aux_distill/mean_u": 0.16614834834143513,
      "aux_distill/n_active_tok": 381.5,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.629968944099379,
      "calib/avg_num_step_conf": 5.89453125,
      "calib/ece": 0.2436470588235294,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.04457453416149071,
      "calib/mean_conf": 0.2167450980392157,
      "calib/mu_c": 0.24121739130434786,
      "calib/mu_w": 0.19664285714285715,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.004705882352941175,
      "calib/std_conf": 0.11159799055349796,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2249.0,
      "completions/max_terminated_length": 2249.0,
      "completions/mean_length": 360.1015625,
      "completions/mean_terminated_length": 360.1015625,
      "completions/min_length": 103.0,
      "completions/min_terminated_length": 103.0,
      "epoch": 0.1888,
      "grad_norm": 0.009426409378647804,
      "learning_rate": 6.388888888888889e-07,
      "loss": 0.0435,
      "num_tokens": 32370660.0,
      "reward": 1.0748529434204102,
      "reward_std": 0.08647443354129791,
      "rewards/accuracy_reward_step": 0.44921875,
      "rewards/final_brier_reward_step": 0.7043933868408203,
      "rewards/format_reward_step": 0.99609375,
      "step": 177
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.383257410954684,
      "aux_distill/mean_u": 0.20492723478918864,
      "aux_distill/n_active_tok": 369.875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.657276119402985,
      "calib/avg_num_step_conf": 5.75390625,
      "calib/ece": 0.2581102362204724,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": 0.06492661691542292,
      "calib/mean_conf": 0.21433070866141732,
      "calib/mu_c": 0.24858333333333338,
      "calib/mu_w": 0.18365671641791045,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.11762137675429811,
      "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": 2004.0,
      "completions/max_terminated_length": 2004.0,
      "completions/mean_length": 360.80859375,
      "completions/mean_terminated_length": 360.80859375,
      "completions/min_length": 95.0,
      "completions/min_terminated_length": 95.0,
      "epoch": 0.18986666666666666,
      "grad_norm": 0.008812584914267063,
      "learning_rate": 6.111111111111112e-07,
      "loss": 0.0495,
      "num_tokens": 32569099.0,
      "reward": 1.0790581703186035,
      "reward_std": 0.10145758092403412,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.6971789002418518,
      "rewards/format_reward_step": 0.9921875,
      "step": 178
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.4004754419438541,
      "aux_distill/mean_u": 0.20431848266051975,
      "aux_distill/n_active_tok": 404.59375,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.6015818858560795,
      "calib/avg_num_step_conf": 6.296875,
      "calib/ece": 0.28051181102362205,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": 0.043978908188585575,
      "calib/mean_conf": 0.2312992125984252,
      "calib/mu_c": 0.25276923076923075,
      "calib/mu_w": 0.20879032258064517,
      "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": 0.11336081444374199,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1880.0,
      "completions/max_terminated_length": 1880.0,
      "completions/mean_length": 385.54296875,
      "completions/mean_terminated_length": 385.54296875,
      "completions/min_length": 129.0,
      "completions/min_terminated_length": 129.0,
      "epoch": 0.19093333333333334,
      "grad_norm": 0.008360245265066624,
      "learning_rate": 5.833333333333334e-07,
      "loss": 0.0576,
      "num_tokens": 32774062.0,
      "reward": 1.087631106376648,
      "reward_std": 0.09885761886835098,
      "rewards/accuracy_reward_step": 0.5078125,
      "rewards/final_brier_reward_step": 0.6752620935440063,
      "rewards/format_reward_step": 0.9921875,
      "step": 179
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33522659935988486,
      "aux_distill/mean_u": 0.20487803825146703,
      "aux_distill/n_active_tok": 378.28125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5125222476481058,
      "calib/avg_num_step_conf": 5.94140625,
      "calib/ece": 0.22277777777777777,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.012578184591914593,
      "calib/mean_conf": 0.22960317460317461,
      "calib/mu_c": 0.23649122807017545,
      "calib/mu_w": 0.22391304347826085,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.10751818446161512,
      "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": 1961.0,
      "completions/max_terminated_length": 1961.0,
      "completions/mean_length": 408.234375,
      "completions/mean_terminated_length": 409.8353271484375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 119.0,
      "epoch": 0.192,
      "grad_norm": 0.008559898473322392,
      "learning_rate": 5.555555555555555e-07,
      "loss": 0.0568,
      "num_tokens": 32982426.0,
      "reward": 1.0580508708953857,
      "reward_std": 0.1284056156873703,
      "rewards/accuracy_reward_step": 0.4453125,
      "rewards/final_brier_reward_step": 0.6864140629768372,
      "rewards/format_reward_step": 0.984375,
      "step": 180
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35764050111174583,
      "aux_distill/mean_u": 0.15735053100971474,
      "aux_distill/n_active_tok": 374.96875,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5122976338729763,
      "calib/avg_num_step_conf": 5.84765625,
      "calib/ece": 0.2119140625,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.00678829389788288,
      "calib/mean_conf": 0.23449218750000003,
      "calib/mu_c": 0.23836363636363628,
      "calib/mu_w": 0.2315753424657534,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.008359375,
      "calib/std_conf": 0.11839311477642965,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1127.0,
      "completions/max_terminated_length": 1127.0,
      "completions/mean_length": 341.95703125,
      "completions/mean_terminated_length": 343.2980651855469,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 106.0,
      "epoch": 0.19306666666666666,
      "grad_norm": 0.010379571467638016,
      "learning_rate": 5.277777777777779e-07,
      "loss": 0.0517,
      "num_tokens": 33176231.0,
      "reward": 1.067920207977295,
      "reward_std": 0.08823342621326447,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.7061527967453003,
      "rewards/format_reward_step": 1.0,
      "step": 181
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3427185930777341,
      "aux_distill/mean_u": 0.1633866363280467,
      "aux_distill/n_active_tok": 423.53125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5718860984271944,
      "calib/avg_num_step_conf": 6.56640625,
      "calib/ece": 0.348543307086614,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": 0.02038305428716386,
      "calib/mean_conf": 0.24208661417322838,
      "calib/mu_c": 0.25075342465753425,
      "calib/mu_w": 0.2303703703703704,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.007913385826771657,
      "calib/std_conf": 0.11763405980218507,
      "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": 2886.0,
      "completions/max_terminated_length": 2886.0,
      "completions/mean_length": 394.640625,
      "completions/mean_terminated_length": 394.640625,
      "completions/min_length": 129.0,
      "completions/min_terminated_length": 129.0,
      "epoch": 0.19413333333333332,
      "grad_norm": 0.008300860412418842,
      "learning_rate": 5.000000000000001e-07,
      "loss": 0.0785,
      "num_tokens": 33383419.0,
      "reward": 1.0954002141952515,
      "reward_std": 0.12059742212295532,
      "rewards/accuracy_reward_step": 0.5703125,
      "rewards/final_brier_reward_step": 0.6322066187858582,
      "rewards/format_reward_step": 0.98828125,
      "step": 182
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.34864079393446445,
      "aux_distill/mean_u": 0.17197240593332896,
      "aux_distill/n_active_tok": 403.40625,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.527304737516005,
      "calib/avg_num_step_conf": 6.28515625,
      "calib/ece": 0.2254761904761905,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.0016542893725992325,
      "calib/mean_conf": 0.23015873015873015,
      "calib/mu_c": 0.23109090909090907,
      "calib/mu_w": 0.22943661971830984,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.009563492063492062,
      "calib/std_conf": 0.10816642179871122,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 1803.0,
      "completions/max_terminated_length": 1803.0,
      "completions/mean_length": 406.1484375,
      "completions/mean_terminated_length": 406.1484375,
      "completions/min_length": 130.0,
      "completions/min_terminated_length": 130.0,
      "epoch": 0.1952,
      "grad_norm": 0.008422174490988255,
      "learning_rate": 4.7222222222222226e-07,
      "loss": 0.0849,
      "num_tokens": 33594073.0,
      "reward": 1.0518405437469482,
      "reward_std": 0.1299438327550888,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.689618706703186,
      "rewards/format_reward_step": 0.984375,
      "step": 183
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3643836765550077,
      "aux_distill/mean_u": 0.1838418861646983,
      "aux_distill/n_active_tok": 372.09375,
      "calib/answer_extract_rate": 1.0,
      "calib/auroc": 0.5901516569930725,
      "calib/avg_num_step_conf": 5.80078125,
      "calib/ece": 0.32921875000000006,
      "calib/final_conf_rate": 1.0,
      "calib/format_rate": 1.0,
      "calib/frac_conf_gt_0.9": 0.00390625,
      "calib/gap": 0.02247768832303565,
      "calib/mean_conf": 0.264375,
      "calib/mu_c": 0.27394557823129256,
      "calib/mu_w": 0.2514678899082569,
      "calib/nonempty_final_conf_rate": 1.0,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.009687500000000003,
      "calib/std_conf": 0.12339754100062124,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 937.0,
      "completions/max_terminated_length": 937.0,
      "completions/mean_length": 345.78515625,
      "completions/mean_terminated_length": 347.1412048339844,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 108.0,
      "epoch": 0.19626666666666667,
      "grad_norm": 0.009029755368828773,
      "learning_rate": 4.444444444444445e-07,
      "loss": 0.0563,
      "num_tokens": 33787874.0,
      "reward": 1.1147441864013672,
      "reward_std": 0.10845861583948135,
      "rewards/accuracy_reward_step": 0.57421875,
      "rewards/final_brier_reward_step": 0.6552695035934448,
      "rewards/format_reward_step": 1.0,
      "step": 184
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.33357510063797235,
      "aux_distill/mean_u": 0.1604589205149373,
      "aux_distill/n_active_tok": 384.25,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5682057965617825,
      "calib/avg_num_step_conf": 5.96875,
      "calib/ece": 0.2504330708661417,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.007874015748031496,
      "calib/gap": 0.025701607397753345,
      "calib/mean_conf": 0.2588582677165354,
      "calib/mu_c": 0.27211382113821136,
      "calib/mu_w": 0.24641221374045802,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.012519685039370083,
      "calib/std_conf": 0.13382573133987627,
      "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": 1285.0,
      "completions/max_terminated_length": 1285.0,
      "completions/mean_length": 363.3515625,
      "completions/mean_terminated_length": 364.7764892578125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 128.0,
      "epoch": 0.19733333333333333,
      "grad_norm": 0.008039743639528751,
      "learning_rate": 4.1666666666666667e-07,
      "loss": 0.026,
      "num_tokens": 33987812.0,
      "reward": 1.0771498680114746,
      "reward_std": 0.12123231589794159,
      "rewards/accuracy_reward_step": 0.48046875,
      "rewards/final_brier_reward_step": 0.6855496168136597,
      "rewards/format_reward_step": 0.98828125,
      "step": 185
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38432842725887895,
      "aux_distill/mean_u": 0.21509393378718752,
      "aux_distill/n_active_tok": 407.28125,
      "calib/answer_extract_rate": 0.98828125,
      "calib/auroc": 0.6162308673469387,
      "calib/avg_num_step_conf": 6.30859375,
      "calib/ece": 0.2046031746031746,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": 0.04655357142857139,
      "calib/mean_conf": 0.24976190476190477,
      "calib/mu_c": 0.27562499999999995,
      "calib/mu_w": 0.22907142857142856,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.004960317460317459,
      "calib/std_conf": 0.12407732027206138,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2598.0,
      "completions/max_terminated_length": 2598.0,
      "completions/mean_length": 380.109375,
      "completions/mean_terminated_length": 381.60003662109375,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 102.0,
      "epoch": 0.1984,
      "grad_norm": 0.008767403662204742,
      "learning_rate": 3.8888888888888895e-07,
      "loss": 0.0441,
      "num_tokens": 34190160.0,
      "reward": 1.0666804313659668,
      "reward_std": 0.12350873649120331,
      "rewards/accuracy_reward_step": 0.4375,
      "rewards/final_brier_reward_step": 0.7114859819412231,
      "rewards/format_reward_step": 0.984375,
      "step": 186
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3616175437346101,
      "aux_distill/mean_u": 0.187725407741795,
      "aux_distill/n_active_tok": 461.90625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5431467181467183,
      "calib/avg_num_step_conf": 7.140625,
      "calib/ece": 0.18505928853754938,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.984375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.018873230373230382,
      "calib/mean_conf": 0.250197628458498,
      "calib/mu_c": 0.26123809523809527,
      "calib/mu_w": 0.2423648648648649,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.010118577075098814,
      "calib/std_conf": 0.1355692159982922,
      "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": 2371.0,
      "completions/max_terminated_length": 2371.0,
      "completions/mean_length": 424.375,
      "completions/mean_terminated_length": 424.375,
      "completions/min_length": 108.0,
      "completions/min_terminated_length": 108.0,
      "epoch": 0.19946666666666665,
      "grad_norm": 0.007748098578304052,
      "learning_rate": 3.611111111111111e-07,
      "loss": 0.058,
      "num_tokens": 34400344.0,
      "reward": 1.0515248775482178,
      "reward_std": 0.1344430148601532,
      "rewards/accuracy_reward_step": 0.41015625,
      "rewards/final_brier_reward_step": 0.708518385887146,
      "rewards/format_reward_step": 0.984375,
      "step": 187
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3558740112930536,
      "aux_distill/mean_u": 0.1870574824077909,
      "aux_distill/n_active_tok": 410.65625,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6338037121789982,
      "calib/avg_num_step_conf": 6.35546875,
      "calib/ece": 0.3025,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.98046875,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": 0.0531464530892449,
      "calib/mean_conf": 0.2505952380952381,
      "calib/mu_c": 0.27463768115942033,
      "calib/mu_w": 0.22149122807017543,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.002738095238095238,
      "calib/std_conf": 0.13092786567064438,
      "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": 2239.0,
      "completions/max_terminated_length": 2239.0,
      "completions/mean_length": 403.42578125,
      "completions/mean_terminated_length": 403.42578125,
      "completions/min_length": 93.0,
      "completions/min_terminated_length": 93.0,
      "epoch": 0.20053333333333334,
      "grad_norm": 0.008264458738267422,
      "learning_rate": 3.3333333333333335e-07,
      "loss": 0.0726,
      "num_tokens": 34607693.0,
      "reward": 1.0906145572662354,
      "reward_std": 0.14208179712295532,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/final_brier_reward_step": 0.661698043346405,
      "rewards/format_reward_step": 0.98046875,
      "step": 188
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3595278114080429,
      "aux_distill/mean_u": 0.18541961370967283,
      "aux_distill/n_active_tok": 444.1875,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5591603053435115,
      "calib/avg_num_step_conf": 6.9140625,
      "calib/ece": 0.2752156862745097,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.00392156862745098,
      "calib/gap": 0.01724082738241814,
      "calib/mean_conf": 0.251921568627451,
      "calib/mu_c": 0.2603053435114504,
      "calib/mu_w": 0.24306451612903227,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.006705882352941177,
      "calib/std_conf": 0.13009134696205646,
      "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": 1585.0,
      "completions/max_terminated_length": 1585.0,
      "completions/mean_length": 411.5078125,
      "completions/mean_terminated_length": 413.12158203125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 114.0,
      "epoch": 0.2016,
      "grad_norm": 0.008612432517111301,
      "learning_rate": 3.055555555555556e-07,
      "loss": 0.0732,
      "num_tokens": 34820807.0,
      "reward": 1.085369348526001,
      "reward_std": 0.10980735719203949,
      "rewards/accuracy_reward_step": 0.51171875,
      "rewards/final_brier_reward_step": 0.666832447052002,
      "rewards/format_reward_step": 0.9921875,
      "step": 189
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3589458088390529,
      "aux_distill/mean_u": 0.1985515963944311,
      "aux_distill/n_active_tok": 439.125,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.6022411616161616,
      "calib/avg_num_step_conf": 6.82421875,
      "calib/ece": 0.20594488188976376,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.98828125,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.03891792929292931,
      "calib/mean_conf": 0.2401181102362205,
      "calib/mu_c": 0.26218181818181824,
      "calib/mu_w": 0.22326388888888893,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.006496062992125984,
      "calib/std_conf": 0.11377378537672232,
      "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": 1861.0,
      "completions/max_terminated_length": 1861.0,
      "completions/mean_length": 416.515625,
      "completions/mean_terminated_length": 418.1490478515625,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 141.0,
      "epoch": 0.20266666666666666,
      "grad_norm": 0.00790832657366991,
      "learning_rate": 2.7777777777777776e-07,
      "loss": 0.0408,
      "num_tokens": 35033043.0,
      "reward": 1.0671627521514893,
      "reward_std": 0.1164865642786026,
      "rewards/accuracy_reward_step": 0.4296875,
      "rewards/final_brier_reward_step": 0.716356635093689,
      "rewards/format_reward_step": 0.98828125,
      "step": 190
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3259891658090055,
      "aux_distill/mean_u": 0.1363302629548188,
      "aux_distill/n_active_tok": 425.53125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5315154922538731,
      "calib/avg_num_step_conf": 6.62890625,
      "calib/ece": 0.202244094488189,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.013333333333333364,
      "calib/mean_conf": 0.277755905511811,
      "calib/mu_c": 0.285,
      "calib/mu_w": 0.2716666666666666,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.011653543307086612,
      "calib/std_conf": 0.1221980589646629,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 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": 2261.0,
      "completions/max_terminated_length": 2261.0,
      "completions/mean_length": 373.3203125,
      "completions/mean_terminated_length": 373.3203125,
      "completions/min_length": 114.0,
      "completions/min_terminated_length": 114.0,
      "epoch": 0.20373333333333332,
      "grad_norm": 0.008117207325994968,
      "learning_rate": 2.5000000000000004e-07,
      "loss": 0.0879,
      "num_tokens": 35232781.0,
      "reward": 1.0756475925445557,
      "reward_std": 0.10618412494659424,
      "rewards/accuracy_reward_step": 0.453125,
      "rewards/final_brier_reward_step": 0.7059824466705322,
      "rewards/format_reward_step": 0.9921875,
      "step": 191
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35070336470380425,
      "aux_distill/mean_u": 0.18195896053579147,
      "aux_distill/n_active_tok": 388.84375,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.645326504481434,
      "calib/avg_num_step_conf": 6.05078125,
      "calib/ece": 0.3147619047619048,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": 0.05792317541613318,
      "calib/mean_conf": 0.2487301587301587,
      "calib/mu_c": 0.27401408450704223,
      "calib/mu_w": 0.21609090909090906,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.0,
      "calib/std_conf": 0.1204487912462106,
      "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": 1481.0,
      "completions/max_terminated_length": 1481.0,
      "completions/mean_length": 373.5,
      "completions/mean_terminated_length": 373.5,
      "completions/min_length": 113.0,
      "completions/min_terminated_length": 113.0,
      "epoch": 0.2048,
      "grad_norm": 0.009182264097034931,
      "learning_rate": 2.2222222222222224e-07,
      "loss": 0.0331,
      "num_tokens": 35433373.0,
      "reward": 1.0928065776824951,
      "reward_std": 0.10991697758436203,
      "rewards/accuracy_reward_step": 0.5546875,
      "rewards/final_brier_reward_step": 0.6543632745742798,
      "rewards/format_reward_step": 0.9765625,
      "step": 192
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.39118539495393634,
      "aux_distill/mean_u": 0.18369763625575194,
      "aux_distill/n_active_tok": 401.9375,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.6368797168236158,
      "calib/avg_num_step_conf": 6.25,
      "calib/ece": 0.12,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9765625,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": 0.06101115340947039,
      "calib/mean_conf": 0.2540157480314961,
      "calib/mu_c": 0.29268817204301073,
      "calib/mu_w": 0.23167701863354034,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9765625,
      "calib/pce": 0.003937007874015748,
      "calib/std_conf": 0.12262714656565303,
      "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": 2157.0,
      "completions/max_terminated_length": 2157.0,
      "completions/mean_length": 391.76953125,
      "completions/mean_terminated_length": 391.76953125,
      "completions/min_length": 103.0,
      "completions/min_terminated_length": 103.0,
      "epoch": 0.20586666666666667,
      "grad_norm": 0.008967787027359009,
      "learning_rate": 1.9444444444444447e-07,
      "loss": 0.05,
      "num_tokens": 35639378.0,
      "reward": 1.043562412261963,
      "reward_std": 0.1039130762219429,
      "rewards/accuracy_reward_step": 0.36328125,
      "rewards/final_brier_reward_step": 0.7472812533378601,
      "rewards/format_reward_step": 0.9765625,
      "step": 193
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3637796104885638,
      "aux_distill/mean_u": 0.18909235012238806,
      "aux_distill/n_active_tok": 324.5,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5746501749125437,
      "calib/avg_num_step_conf": 5.05859375,
      "calib/ece": 0.2801181102362204,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.030187406296851516,
      "calib/mean_conf": 0.2729527559055119,
      "calib/mu_c": 0.28673913043478255,
      "calib/mu_w": 0.25655172413793104,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.9921875,
      "calib/nonempty_step_conf_rate": 0.9921875,
      "calib/pce": 0.004881889763779529,
      "calib/std_conf": 0.12136178478979952,
      "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": 2007.0,
      "completions/max_terminated_length": 2007.0,
      "completions/mean_length": 339.18359375,
      "completions/mean_terminated_length": 339.18359375,
      "completions/min_length": 122.0,
      "completions/min_terminated_length": 122.0,
      "epoch": 0.20693333333333333,
      "grad_norm": 0.009117648005485535,
      "learning_rate": 1.6666666666666668e-07,
      "loss": 0.066,
      "num_tokens": 35832153.0,
      "reward": 1.1024904251098633,
      "reward_std": 0.11599205434322357,
      "rewards/accuracy_reward_step": 0.5390625,
      "rewards/final_brier_reward_step": 0.6737308502197266,
      "rewards/format_reward_step": 0.9921875,
      "step": 194
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3681975821964443,
      "aux_distill/mean_u": 0.21568309308917685,
      "aux_distill/n_active_tok": 384.53125,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.5918079096045198,
      "calib/avg_num_step_conf": 6.0546875,
      "calib/ece": 0.2940316205533597,
      "calib/final_conf_rate": 0.98828125,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.03751851851851845,
      "calib/mean_conf": 0.24501976284584978,
      "calib/mu_c": 0.2625185185185185,
      "calib/mu_w": 0.22500000000000003,
      "calib/nonempty_final_conf_rate": 0.98828125,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.002727272727272727,
      "calib/std_conf": 0.11820001974923042,
      "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": 2956.0,
      "completions/max_terminated_length": 2956.0,
      "completions/mean_length": 386.69921875,
      "completions/mean_terminated_length": 388.2156982421875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 95.0,
      "epoch": 0.208,
      "grad_norm": 0.009348858147859573,
      "learning_rate": 1.3888888888888888e-07,
      "loss": 0.0937,
      "num_tokens": 36037132.0,
      "reward": 1.0756480693817139,
      "reward_std": 0.1506083607673645,
      "rewards/accuracy_reward_step": 0.52734375,
      "rewards/final_brier_reward_step": 0.651296079158783,
      "rewards/format_reward_step": 0.97265625,
      "step": 195
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.3619554964825511,
      "aux_distill/mean_u": 0.16458977785492765,
      "aux_distill/n_active_tok": 369.5625,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.6130852713178295,
      "calib/avg_num_step_conf": 5.73046875,
      "calib/ece": 0.20559055118110234,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": 0.050297054263565955,
      "calib/mean_conf": 0.28653543307086615,
      "calib/mu_c": 0.31208,
      "calib/mu_w": 0.26178294573643407,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0,
      "calib/std_conf": 0.11713023878992854,
      "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": 2781.0,
      "completions/max_terminated_length": 2781.0,
      "completions/mean_length": 337.21484375,
      "completions/mean_terminated_length": 337.21484375,
      "completions/min_length": 97.0,
      "completions/min_terminated_length": 97.0,
      "epoch": 0.20906666666666668,
      "grad_norm": 0.008571833372116089,
      "learning_rate": 1.1111111111111112e-07,
      "loss": 0.0933,
      "num_tokens": 36226003.0,
      "reward": 1.0970335006713867,
      "reward_std": 0.11255770921707153,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/final_brier_reward_step": 0.7135984897613525,
      "rewards/format_reward_step": 0.9921875,
      "step": 196
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38431750517338514,
      "aux_distill/mean_u": 0.23451896721609689,
      "aux_distill/n_active_tok": 441.875,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5431704260651631,
      "calib/avg_num_step_conf": 6.859375,
      "calib/ece": 0.19200787401574806,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": 0.022473684210526312,
      "calib/mean_conf": 0.26208661417322837,
      "calib/mu_c": 0.2744736842105263,
      "calib/mu_w": 0.252,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.99609375,
      "calib/pce": 0.0026377952755905513,
      "calib/std_conf": 0.11638576391162316,
      "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": 2664.0,
      "completions/max_terminated_length": 2664.0,
      "completions/mean_length": 381.55859375,
      "completions/mean_terminated_length": 384.56298828125,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 131.0,
      "epoch": 0.21013333333333334,
      "grad_norm": 0.008849774487316608,
      "learning_rate": 8.333333333333334e-08,
      "loss": -0.0048,
      "num_tokens": 36428738.0,
      "reward": 1.0736178159713745,
      "reward_std": 0.1103426069021225,
      "rewards/accuracy_reward_step": 0.4453125,
      "rewards/final_brier_reward_step": 0.7097355127334595,
      "rewards/format_reward_step": 0.9921875,
      "step": 197
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.38216920103877783,
      "aux_distill/mean_u": 0.17609322742804628,
      "aux_distill/n_active_tok": 426.65625,
      "calib/answer_extract_rate": 0.99609375,
      "calib/auroc": 0.5772788947822869,
      "calib/avg_num_step_conf": 6.61328125,
      "calib/ece": 0.2735686274509804,
      "calib/final_conf_rate": 0.99609375,
      "calib/format_rate": 0.99609375,
      "calib/frac_conf_gt_0.9": 0.0,
      "calib/gap": 0.027984457875909724,
      "calib/mean_conf": 0.2570196078431372,
      "calib/mu_c": 0.2702985074626866,
      "calib/mu_w": 0.24231404958677685,
      "calib/nonempty_final_conf_rate": 0.99609375,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.0025490196078431387,
      "calib/std_conf": 0.12130555977043575,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 1590.0,
      "completions/max_terminated_length": 1590.0,
      "completions/mean_length": 379.24609375,
      "completions/mean_terminated_length": 380.7333679199219,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 92.0,
      "epoch": 0.2112,
      "grad_norm": 0.008252741768956184,
      "learning_rate": 5.555555555555556e-08,
      "loss": 0.0495,
      "num_tokens": 36631209.0,
      "reward": 1.097348928451538,
      "reward_std": 0.10788056254386902,
      "rewards/accuracy_reward_step": 0.5234375,
      "rewards/final_brier_reward_step": 0.6751663684844971,
      "rewards/format_reward_step": 0.99609375,
      "step": 198
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.35966091603040695,
      "aux_distill/mean_u": 0.20198334396455808,
      "aux_distill/n_active_tok": 390.125,
      "calib/answer_extract_rate": 0.9921875,
      "calib/auroc": 0.5756899224806202,
      "calib/avg_num_step_conf": 6.02734375,
      "calib/ece": 0.2374803149606299,
      "calib/final_conf_rate": 0.9921875,
      "calib/format_rate": 0.9921875,
      "calib/frac_conf_gt_0.9": 0.003937007874015748,
      "calib/gap": 0.034306976744186146,
      "calib/mean_conf": 0.2689763779527559,
      "calib/mu_c": 0.28640000000000004,
      "calib/mu_w": 0.2520930232558139,
      "calib/nonempty_final_conf_rate": 0.9921875,
      "calib/nonempty_reasoning_rate": 1.0,
      "calib/nonempty_step_conf_rate": 1.0,
      "calib/pce": 0.007165354330708661,
      "calib/std_conf": 0.12804917021667883,
      "calib/step_conf_rate": 1.0,
      "clip_ratio/high_max": 0.0,
      "clip_ratio/high_mean": 0.0,
      "clip_ratio/low_mean": 0.0,
      "clip_ratio/low_min": 0.0,
      "clip_ratio/region_mean": 0.0,
      "completions/clipped_ratio": 0.00390625,
      "completions/max_length": 2242.0,
      "completions/max_terminated_length": 2242.0,
      "completions/mean_length": 392.3828125,
      "completions/mean_terminated_length": 393.9216003417969,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 96.0,
      "epoch": 0.21226666666666666,
      "grad_norm": 0.008318433538079262,
      "learning_rate": 2.777777777777778e-08,
      "loss": 0.0254,
      "num_tokens": 36835859.0,
      "reward": 1.0880054235458374,
      "reward_std": 0.11319223791360855,
      "rewards/accuracy_reward_step": 0.48828125,
      "rewards/final_brier_reward_step": 0.6955421566963196,
      "rewards/format_reward_step": 0.9921875,
      "step": 199
    },
    {
      "aux_distill/lambda": 0.10000000000000005,
      "aux_distill/loss": 0.36204342264682055,
      "aux_distill/mean_u": 0.191237035896235,
      "aux_distill/n_active_tok": 377.15625,
      "calib/answer_extract_rate": 0.984375,
      "calib/auroc": 0.6301767676767677,
      "calib/avg_num_step_conf": 5.80859375,
      "calib/ece": 0.22722222222222221,
      "calib/final_conf_rate": 0.984375,
      "calib/format_rate": 0.97265625,
      "calib/frac_conf_gt_0.9": 0.003968253968253968,
      "calib/gap": 0.05615909090909091,
      "calib/mean_conf": 0.25166666666666665,
      "calib/mu_c": 0.28108333333333335,
      "calib/mu_w": 0.22492424242424244,
      "calib/nonempty_final_conf_rate": 0.984375,
      "calib/nonempty_reasoning_rate": 0.99609375,
      "calib/nonempty_step_conf_rate": 0.984375,
      "calib/pce": 0.0013492063492063487,
      "calib/std_conf": 0.1274521385424647,
      "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": 2290.0,
      "completions/max_terminated_length": 2290.0,
      "completions/mean_length": 396.66015625,
      "completions/mean_terminated_length": 398.2156982421875,
      "completions/min_length": 0.0,
      "completions/min_terminated_length": 81.0,
      "epoch": 0.21333333333333335,
      "grad_norm": 0.009059817530214787,
      "learning_rate": 0.0,
      "loss": 0.0259,
      "num_tokens": 37045452.0,
      "reward": 1.0666258335113525,
      "reward_std": 0.12554016709327698,
      "rewards/accuracy_reward_step": 0.46875,
      "rewards/final_brier_reward_step": 0.6918452978134155,
      "rewards/format_reward_step": 0.97265625,
      "step": 200
    },
    {
      "epoch": 0.21333333333333335,
      "step": 200,
      "total_flos": 0.0,
      "train_loss": 0.06970496780704707,
      "train_runtime": 12590.6678,
      "train_samples_per_second": 4.067,
      "train_steps_per_second": 0.016
    }
  ],
  "logging_steps": 1,
  "max_steps": 200,
  "num_input_tokens_seen": 37045452,
  "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
}