{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.21333333333333335, "eval_steps": 500, "global_step": 200, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "calib/answer_extract_rate": 0.04296875, "calib/auroc": 0.75, "calib/avg_num_step_conf": 0.09375, "calib/ece": 0.2866666666666666, "calib/final_conf_rate": 0.0234375, "calib/format_rate": 0.0234375, "calib/frac_conf_gt_0.9": 0.16666666666666666, "calib/gap": 0.022499999999999964, "calib/mean_conf": 0.85, "calib/mu_c": 0.8574999999999999, "calib/mu_w": 0.835, "calib/nonempty_final_conf_rate": 0.0234375, "calib/nonempty_reasoning_rate": 0.046875, "calib/nonempty_step_conf_rate": 0.02734375, "calib/pce": 0.23499999999999993, "calib/std_conf": 0.0912870929175277, "calib/step_conf_rate": 0.02734375, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 3000.0, "completions/max_terminated_length": 3000.0, "completions/mean_length": 625.296875, "completions/mean_terminated_length": 695.9826049804688, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.0010666666666666667, "grad_norm": 0.0033870742190629244, "learning_rate": 2.5000000000000004e-07, "loss": -0.0038, "num_tokens": 321164.0, "reward": 0.044097598642110825, "reward_std": 0.10592999309301376, "rewards/accuracy_reward_step": 0.015625, "rewards/final_brier_reward_step": 0.01748046837747097, "rewards/format_reward_step": 0.0234375, "rewards/stepwise_brier_reward": 0.01828492246568203, "step": 1 }, { "calib/answer_extract_rate": 0.0390625, "calib/auroc": 0.7857142857142857, "calib/avg_num_step_conf": 0.25, "calib/ece": 0.6511111111111112, "calib/final_conf_rate": 0.03515625, "calib/format_rate": 0.03125, "calib/frac_conf_gt_0.9": 0.2222222222222222, "calib/gap": 0.0535714285714286, "calib/mean_conf": 0.8733333333333334, "calib/mu_c": 0.915, "calib/mu_w": 0.8614285714285714, "calib/nonempty_final_conf_rate": 0.03515625, "calib/nonempty_reasoning_rate": 0.06640625, "calib/nonempty_step_conf_rate": 0.05859375, "calib/pce": 0.6511111111111112, "calib/std_conf": 0.05270462766947297, "calib/step_conf_rate": 0.05859375, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.05859375, "completions/max_length": 2779.0, "completions/max_terminated_length": 2779.0, "completions/mean_length": 629.30078125, "completions/mean_terminated_length": 668.4689331054688, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.0021333333333333334, "grad_norm": 0.0027943712193518877, "learning_rate": 5.000000000000001e-07, "loss": 0.0139, "num_tokens": 639057.0, "reward": 0.035815127193927765, "reward_std": 0.06483516097068787, "rewards/accuracy_reward_step": 0.0078125, "rewards/final_brier_reward_step": 0.0137117188423872, "rewards/format_reward_step": 0.03125, "rewards/stepwise_brier_reward": 0.02017379179596901, "step": 2 }, { "calib/answer_extract_rate": 0.06640625, "calib/auroc": 0.7291666666666666, "calib/avg_num_step_conf": 0.296875, "calib/ece": 0.5209090909090909, "calib/final_conf_rate": 0.04296875, "calib/format_rate": 0.02734375, "calib/frac_conf_gt_0.9": 0.36363636363636365, "calib/gap": 0.18291666666666673, "calib/mean_conf": 0.7936363636363637, "calib/mu_c": 0.9266666666666667, "calib/mu_w": 0.74375, "calib/nonempty_final_conf_rate": 0.04296875, "calib/nonempty_reasoning_rate": 0.08203125, "calib/nonempty_step_conf_rate": 0.0546875, "calib/pce": 0.5209090909090909, "calib/std_conf": 0.27360464932495465, "calib/step_conf_rate": 0.0546875, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0859375, "completions/max_length": 2970.0, "completions/max_terminated_length": 2970.0, "completions/mean_length": 655.51953125, "completions/mean_terminated_length": 717.1495971679688, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.0032, "grad_norm": 0.003702443325892091, "learning_rate": 7.5e-07, "loss": 0.0069, "num_tokens": 965630.0, "reward": 0.03842926025390625, "reward_std": 0.08950819075107574, "rewards/accuracy_reward_step": 0.01171875, "rewards/final_brier_reward_step": 0.01211836002767086, "rewards/format_reward_step": 0.02734375, "rewards/stepwise_brier_reward": 0.016598684713244438, "step": 3 }, { "calib/answer_extract_rate": 0.02734375, "calib/avg_num_step_conf": 0.10546875, "calib/ece": 0.8, "calib/final_conf_rate": 0.00390625, "calib/format_rate": 0.0, "calib/frac_conf_gt_0.9": 0.0, "calib/mean_conf": 0.8, "calib/mu_c": NaN, "calib/mu_w": 0.8, "calib/nonempty_final_conf_rate": 0.00390625, "calib/nonempty_reasoning_rate": 0.0390625, "calib/nonempty_step_conf_rate": 0.015625, "calib/pce": 0.8, "calib/std_conf": 0.0, "calib/step_conf_rate": 0.015625, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.125, "completions/max_length": 2938.0, "completions/max_terminated_length": 2938.0, "completions/mean_length": 655.375, "completions/mean_terminated_length": 749.0000610351562, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.004266666666666667, "grad_norm": 0.0, "learning_rate": 1.0000000000000002e-06, "loss": 0.0, "num_tokens": 1293078.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_step": 0.0, "rewards/final_brier_reward_step": 0.0, "rewards/format_reward_step": 0.0, "rewards/stepwise_brier_reward": 0.0, "step": 4 }, { "calib/answer_extract_rate": 0.04296875, "calib/avg_num_step_conf": 0.1484375, "calib/ece": 0.8025, "calib/final_conf_rate": 0.03125, "calib/format_rate": 0.01171875, "calib/frac_conf_gt_0.9": 0.25, "calib/mean_conf": 0.8025, "calib/mu_c": NaN, "calib/mu_w": 0.8025, "calib/nonempty_final_conf_rate": 0.03125, "calib/nonempty_reasoning_rate": 0.05078125, "calib/nonempty_step_conf_rate": 0.02734375, "calib/pce": 0.8025, "calib/std_conf": 0.14964541422977184, "calib/step_conf_rate": 0.02734375, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.1484375, "completions/max_length": 3014.0, "completions/max_terminated_length": 3014.0, "completions/mean_length": 660.1171875, "completions/mean_terminated_length": 775.1834716796875, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.005333333333333333, "grad_norm": 0.0016447411617264152, "learning_rate": 1.25e-06, "loss": 0.0028, "num_tokens": 1622260.0, "reward": 0.00818483904004097, "reward_std": 0.017666654661297798, "rewards/accuracy_reward_step": 0.0, "rewards/final_brier_reward_step": 0.003939843736588955, "rewards/format_reward_step": 0.01171875, "rewards/stepwise_brier_reward": 0.00536201149225235, "step": 5 }, { "calib/answer_extract_rate": 0.0546875, "calib/auroc": 0.375, "calib/avg_num_step_conf": 0.1484375, "calib/ece": 0.7322222222222222, "calib/final_conf_rate": 0.03515625, "calib/format_rate": 0.015625, "calib/frac_conf_gt_0.9": 0.4444444444444444, "calib/gap": 0.0012499999999999734, "calib/mean_conf": 0.798888888888889, "calib/mu_c": 0.8, "calib/mu_w": 0.7987500000000001, "calib/nonempty_final_conf_rate": 0.03515625, "calib/nonempty_reasoning_rate": 0.0703125, "calib/nonempty_step_conf_rate": 0.046875, "calib/pce": 0.71, "calib/std_conf": 0.22382753501769787, "calib/step_conf_rate": 0.046875, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.08203125, "completions/max_length": 2993.0, "completions/max_terminated_length": 2993.0, "completions/mean_length": 653.5234375, "completions/mean_terminated_length": 711.9234008789062, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.0064, "grad_norm": 0.0018285529222339392, "learning_rate": 1.5e-06, "loss": 0.0195, "num_tokens": 1949018.0, "reward": 0.016980905085802078, "reward_std": 0.03552158549427986, "rewards/accuracy_reward_step": 0.00390625, "rewards/final_brier_reward_step": 0.003566406201571226, "rewards/format_reward_step": 0.015625, "rewards/stepwise_brier_reward": 0.009669709019362926, "step": 6 }, { "calib/answer_extract_rate": 0.0625, "calib/auroc": 0.5625, "calib/avg_num_step_conf": 0.10546875, "calib/ece": 0.5980000000000001, "calib/final_conf_rate": 0.0390625, "calib/format_rate": 0.02734375, "calib/frac_conf_gt_0.9": 0.4, "calib/gap": 0.09624999999999995, "calib/mean_conf": 0.798, "calib/mu_c": 0.875, "calib/mu_w": 0.77875, "calib/nonempty_final_conf_rate": 0.0390625, "calib/nonempty_reasoning_rate": 0.0703125, "calib/nonempty_step_conf_rate": 0.0390625, "calib/pce": 0.5980000000000001, "calib/std_conf": 0.21761433776293326, "calib/step_conf_rate": 0.0390625, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.08984375, "completions/max_length": 2981.0, "completions/max_terminated_length": 2981.0, "completions/mean_length": 649.796875, "completions/mean_terminated_length": 713.93994140625, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.007466666666666667, "grad_norm": 0.004392486996948719, "learning_rate": 1.75e-06, "loss": -0.0027, "num_tokens": 2276294.0, "reward": 0.03461461514234543, "reward_std": 0.07292924076318741, "rewards/accuracy_reward_step": 0.0078125, "rewards/final_brier_reward_step": 0.01576288975775242, "rewards/format_reward_step": 0.02734375, "rewards/stepwise_brier_reward": 0.021133072674274445, "step": 7 }, { "calib/answer_extract_rate": 0.046875, "calib/avg_num_step_conf": 0.1875, "calib/ece": 0.90625, "calib/final_conf_rate": 0.03125, "calib/format_rate": 0.0234375, "calib/frac_conf_gt_0.9": 0.375, "calib/mean_conf": 0.90625, "calib/mu_c": NaN, "calib/mu_w": 0.90625, "calib/nonempty_final_conf_rate": 0.03125, "calib/nonempty_reasoning_rate": 0.05859375, "calib/nonempty_step_conf_rate": 0.04296875, "calib/pce": 0.90625, "calib/std_conf": 0.04240798863421843, "calib/step_conf_rate": 0.04296875, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0859375, "completions/max_length": 3052.0, "completions/max_terminated_length": 3052.0, "completions/mean_length": 680.63671875, "completions/mean_terminated_length": 744.6282348632812, "completions/min_length": 0.0, "completions/min_terminated_length": 3.0, "epoch": 0.008533333333333334, "grad_norm": 0.003672001650556922, "learning_rate": 2.0000000000000003e-06, "loss": 0.0249, "num_tokens": 2610553.0, "reward": 0.01645149476826191, "reward_std": 0.04114970937371254, "rewards/accuracy_reward_step": 0.0, "rewards/final_brier_reward_step": 0.004266406409442425, "rewards/format_reward_step": 0.0234375, "rewards/stepwise_brier_reward": 0.014664572663605213, "step": 8 }, { "calib/answer_extract_rate": 0.04296875, "calib/auroc": 1.0, "calib/avg_num_step_conf": 0.11328125, "calib/ece": 0.202, "calib/final_conf_rate": 0.01953125, "calib/format_rate": 0.015625, "calib/frac_conf_gt_0.9": 0.4, "calib/gap": 0.55, "calib/mean_conf": 0.6699999999999999, "calib/mu_c": 0.89, "calib/mu_w": 0.34, "calib/nonempty_final_conf_rate": 0.01953125, "calib/nonempty_reasoning_rate": 0.04296875, "calib/nonempty_step_conf_rate": 0.01953125, "calib/pce": 0.136, "calib/std_conf": 0.3011312006418465, "calib/step_conf_rate": 0.01953125, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.12109375, "completions/max_length": 3066.0, "completions/max_terminated_length": 3066.0, "completions/mean_length": 543.2578125, "completions/mean_terminated_length": 618.106689453125, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.0096, "grad_norm": 0.0025936109013855457, "learning_rate": 2.25e-06, "loss": 0.0123, "num_tokens": 2910667.0, "reward": 0.03152869641780853, "reward_std": 0.08917662501335144, "rewards/accuracy_reward_step": 0.01171875, "rewards/final_brier_reward_step": 0.015242968685925007, "rewards/format_reward_step": 0.015625, "rewards/stepwise_brier_reward": 0.009309324435889721, "step": 9 }, { "calib/answer_extract_rate": 0.05859375, "calib/auroc": 0.625, "calib/avg_num_step_conf": 0.15234375, "calib/ece": 0.6037500000000001, "calib/final_conf_rate": 0.03125, "calib/format_rate": 0.01953125, "calib/frac_conf_gt_0.9": 0.25, "calib/gap": 0.02833333333333332, "calib/mean_conf": 0.85375, "calib/mu_c": 0.875, "calib/mu_w": 0.8466666666666667, "calib/nonempty_final_conf_rate": 0.03125, "calib/nonempty_reasoning_rate": 0.06640625, "calib/nonempty_step_conf_rate": 0.0390625, "calib/pce": 0.6037500000000001, "calib/std_conf": 0.07016721100343094, "calib/step_conf_rate": 0.0390625, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 2947.0, "completions/max_terminated_length": 2947.0, "completions/mean_length": 750.6484375, "completions/mean_terminated_length": 800.6917114257812, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.010666666666666666, "grad_norm": 0.002753326902166009, "learning_rate": 2.5e-06, "loss": 0.0024, "num_tokens": 3263137.0, "reward": 0.027742739766836166, "reward_std": 0.07846832275390625, "rewards/accuracy_reward_step": 0.0078125, "rewards/final_brier_reward_step": 0.011815235018730164, "rewards/format_reward_step": 0.01953125, "rewards/stepwise_brier_reward": 0.0132182277739048, "step": 10 }, { "calib/answer_extract_rate": 0.1015625, "calib/auroc": 0.725, "calib/avg_num_step_conf": 0.17578125, "calib/ece": 0.74, "calib/final_conf_rate": 0.046875, "calib/format_rate": 0.0390625, "calib/frac_conf_gt_0.9": 0.5833333333333334, "calib/gap": 0.05800000000000005, "calib/mean_conf": 0.9066666666666667, "calib/mu_c": 0.9550000000000001, "calib/mu_w": 0.897, "calib/nonempty_final_conf_rate": 0.046875, "calib/nonempty_reasoning_rate": 0.109375, "calib/nonempty_step_conf_rate": 0.0546875, "calib/pce": 0.74, "calib/std_conf": 0.07575545451575674, "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": 0.109375, "completions/max_length": 3029.0, "completions/max_terminated_length": 3029.0, "completions/mean_length": 618.66015625, "completions/mean_terminated_length": 694.635986328125, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.011733333333333333, "grad_norm": 0.0041241636499762535, "learning_rate": 2.7500000000000004e-06, "loss": 0.0152, "num_tokens": 3579498.0, "reward": 0.04076514020562172, "reward_std": 0.10268531739711761, "rewards/accuracy_reward_step": 0.0078125, "rewards/final_brier_reward_step": 0.012591796927154064, "rewards/format_reward_step": 0.0390625, "rewards/stepwise_brier_reward": 0.02546875737607479, "step": 11 }, { "calib/answer_extract_rate": 0.078125, "calib/auroc": 0.5416666666666667, "calib/avg_num_step_conf": 0.1640625, "calib/ece": 0.525, "calib/final_conf_rate": 0.05078125, "calib/format_rate": 0.03515625, "calib/frac_conf_gt_0.9": 0.23076923076923078, "calib/gap": 0.036111111111111094, "calib/mean_conf": 0.7250000000000002, "calib/mu_c": 0.75, "calib/mu_w": 0.7138888888888889, "calib/nonempty_final_conf_rate": 0.05078125, "calib/nonempty_reasoning_rate": 0.08984375, "calib/nonempty_step_conf_rate": 0.0546875, "calib/pce": 0.4711538461538462, "calib/std_conf": 0.28774989139876317, "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": 0.08984375, "completions/max_length": 3067.0, "completions/max_terminated_length": 3067.0, "completions/mean_length": 674.59375, "completions/mean_terminated_length": 741.1845703125, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.0128, "grad_norm": 0.00357402884401381, "learning_rate": 3e-06, "loss": 0.0251, "num_tokens": 3909874.0, "reward": 0.06389623135328293, "reward_std": 0.12676812708377838, "rewards/accuracy_reward_step": 0.0234375, "rewards/final_brier_reward_step": 0.02101787179708481, "rewards/format_reward_step": 0.03515625, "rewards/stepwise_brier_reward": 0.02362954430282116, "step": 12 }, { "calib/answer_extract_rate": 0.0859375, "calib/auroc": 0.5625, "calib/avg_num_step_conf": 0.25390625, "calib/ece": 0.5220000000000001, "calib/final_conf_rate": 0.0390625, "calib/format_rate": 0.02734375, "calib/frac_conf_gt_0.9": 0.4, "calib/gap": -0.020000000000000018, "calib/mean_conf": 0.8220000000000001, "calib/mu_c": 0.8099999999999999, "calib/mu_w": 0.83, "calib/nonempty_final_conf_rate": 0.0390625, "calib/nonempty_reasoning_rate": 0.11328125, "calib/nonempty_step_conf_rate": 0.07421875, "calib/pce": 0.4720000000000001, "calib/std_conf": 0.13444701558606648, "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": 0.05859375, "completions/max_length": 3049.0, "completions/max_terminated_length": 3049.0, "completions/mean_length": 627.3046875, "completions/mean_terminated_length": 666.3485717773438, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.013866666666666666, "grad_norm": 0.003709837095811963, "learning_rate": 3.2500000000000002e-06, "loss": 0.0205, "num_tokens": 4228560.0, "reward": 0.04722949117422104, "reward_std": 0.12521444261074066, "rewards/accuracy_reward_step": 0.015625, "rewards/final_brier_reward_step": 0.01871640607714653, "rewards/format_reward_step": 0.02734375, "rewards/stepwise_brier_reward": 0.021764058619737625, "step": 13 }, { "calib/answer_extract_rate": 0.1015625, "calib/auroc": 0.5113636363636364, "calib/avg_num_step_conf": 0.2265625, "calib/ece": 0.5254933333333334, "calib/final_conf_rate": 0.05859375, "calib/format_rate": 0.04296875, "calib/frac_conf_gt_0.9": 0.4, "calib/gap": 0.1436454545454544, "calib/mean_conf": 0.7921599999999999, "calib/mu_c": 0.8975, "calib/mu_w": 0.7538545454545456, "calib/nonempty_final_conf_rate": 0.05859375, "calib/nonempty_reasoning_rate": 0.1171875, "calib/nonempty_step_conf_rate": 0.06640625, "calib/pce": 0.5254933333333334, "calib/std_conf": 0.2851773455237986, "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": 0.0703125, "completions/max_length": 2960.0, "completions/max_terminated_length": 2960.0, "completions/mean_length": 646.53125, "completions/mean_terminated_length": 695.4285888671875, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.014933333333333333, "grad_norm": 0.00326512916944921, "learning_rate": 3.5e-06, "loss": 0.0289, "num_tokens": 4552976.0, "reward": 0.058872662484645844, "reward_std": 0.12486343085765839, "rewards/accuracy_reward_step": 0.015625, "rewards/final_brier_reward_step": 0.019893750548362732, "rewards/format_reward_step": 0.04296875, "rewards/stepwise_brier_reward": 0.035909418016672134, "step": 14 }, { "calib/answer_extract_rate": 0.09375, "calib/auroc": 0.75, "calib/avg_num_step_conf": 0.2265625, "calib/ece": 0.5292307692307692, "calib/final_conf_rate": 0.05078125, "calib/format_rate": 0.0390625, "calib/frac_conf_gt_0.9": 0.3076923076923077, "calib/gap": 0.12277777777777776, "calib/mean_conf": 0.7600000000000002, "calib/mu_c": 0.845, "calib/mu_w": 0.7222222222222222, "calib/nonempty_final_conf_rate": 0.05078125, "calib/nonempty_reasoning_rate": 0.11328125, "calib/nonempty_step_conf_rate": 0.05859375, "calib/pce": 0.4907692307692308, "calib/std_conf": 0.19074751737471024, "calib/step_conf_rate": 0.05859375, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 2924.0, "completions/max_terminated_length": 2924.0, "completions/mean_length": 586.4765625, "completions/mean_terminated_length": 625.5750122070312, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.016, "grad_norm": 0.004567032679915428, "learning_rate": 3.7500000000000005e-06, "loss": 0.0067, "num_tokens": 4864498.0, "reward": 0.06275875866413116, "reward_std": 0.1310034841299057, "rewards/accuracy_reward_step": 0.01953125, "rewards/final_brier_reward_step": 0.023592187091708183, "rewards/format_reward_step": 0.0390625, "rewards/stepwise_brier_reward": 0.03213036060333252, "step": 15 }, { "calib/answer_extract_rate": 0.16796875, "calib/auroc": 0.5684523809523809, "calib/avg_num_step_conf": 0.75, "calib/ece": 0.6538235294117647, "calib/final_conf_rate": 0.1328125, "calib/format_rate": 0.09375, "calib/frac_conf_gt_0.9": 0.29411764705882354, "calib/gap": -0.0013095238095238715, "calib/mean_conf": 0.7844117647058823, "calib/mu_c": 0.7833333333333333, "calib/mu_w": 0.7846428571428572, "calib/nonempty_final_conf_rate": 0.1328125, "calib/nonempty_reasoning_rate": 0.203125, "calib/nonempty_step_conf_rate": 0.140625, "calib/pce": 0.6308823529411764, "calib/std_conf": 0.2636083455841103, "calib/step_conf_rate": 0.140625, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0859375, "completions/max_length": 2935.0, "completions/max_terminated_length": 2935.0, "completions/mean_length": 654.53515625, "completions/mean_terminated_length": 716.0726928710938, "completions/min_length": 0.0, "completions/min_terminated_length": 7.0, "epoch": 0.017066666666666667, "grad_norm": 0.0057198419235646725, "learning_rate": 4.000000000000001e-06, "loss": 0.0684, "num_tokens": 5194411.0, "reward": 0.11228662729263306, "reward_std": 0.2298872321844101, "rewards/accuracy_reward_step": 0.02734375, "rewards/final_brier_reward_step": 0.03472773730754852, "rewards/format_reward_step": 0.09375, "rewards/stepwise_brier_reward": 0.0628562718629837, "step": 16 }, { "calib/answer_extract_rate": 0.2109375, "calib/auroc": 0.6428571428571429, "calib/avg_num_step_conf": 0.66796875, "calib/ece": 0.5893636363636363, "calib/final_conf_rate": 0.171875, "calib/format_rate": 0.14453125, "calib/frac_conf_gt_0.9": 0.38636363636363635, "calib/gap": 0.10397460317460328, "calib/mean_conf": 0.7761818181818182, "calib/mu_c": 0.8588888888888889, "calib/mu_w": 0.7549142857142856, "calib/nonempty_final_conf_rate": 0.171875, "calib/nonempty_reasoning_rate": 0.25, "calib/nonempty_step_conf_rate": 0.19140625, "calib/pce": 0.5805, "calib/std_conf": 0.27200361567844783, "calib/step_conf_rate": 0.19140625, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.07421875, "completions/max_length": 3057.0, "completions/max_terminated_length": 3057.0, "completions/mean_length": 515.56640625, "completions/mean_terminated_length": 556.898681640625, "completions/min_length": 0.0, "completions/min_terminated_length": 3.0, "epoch": 0.018133333333333335, "grad_norm": 0.008250598795711994, "learning_rate": 4.25e-06, "loss": 0.0284, "num_tokens": 5483428.0, "reward": 0.16988542675971985, "reward_std": 0.3009862005710602, "rewards/accuracy_reward_step": 0.03515625, "rewards/final_brier_reward_step": 0.07160412520170212, "rewards/format_reward_step": 0.14453125, "rewards/stepwise_brier_reward": 0.10793761909008026, "step": 17 }, { "calib/answer_extract_rate": 0.20703125, "calib/auroc": 0.5384615384615385, "calib/avg_num_step_conf": 0.56640625, "calib/ece": 0.6722727272727272, "calib/final_conf_rate": 0.171875, "calib/format_rate": 0.109375, "calib/frac_conf_gt_0.9": 0.4318181818181818, "calib/gap": 0.0700512820512822, "calib/mean_conf": 0.785909090909091, "calib/mu_c": 0.8480000000000001, "calib/mu_w": 0.7779487179487179, "calib/nonempty_final_conf_rate": 0.171875, "calib/nonempty_reasoning_rate": 0.2421875, "calib/nonempty_step_conf_rate": 0.15625, "calib/pce": 0.6722727272727272, "calib/std_conf": 0.2645950437746003, "calib/step_conf_rate": 0.15625, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0625, "completions/max_length": 2959.0, "completions/max_terminated_length": 2959.0, "completions/mean_length": 590.5390625, "completions/mean_terminated_length": 629.9083862304688, "completions/min_length": 0.0, "completions/min_terminated_length": 3.0, "epoch": 0.0192, "grad_norm": 0.006911951582878828, "learning_rate": 4.5e-06, "loss": 0.0755, "num_tokens": 5798830.0, "reward": 0.1128019392490387, "reward_std": 0.2207101732492447, "rewards/accuracy_reward_step": 0.01953125, "rewards/final_brier_reward_step": 0.046952731907367706, "rewards/format_reward_step": 0.109375, "rewards/stepwise_brier_reward": 0.06831751763820648, "step": 18 }, { "calib/answer_extract_rate": 0.49609375, "calib/auroc": 0.5732779677576941, "calib/avg_num_step_conf": 1.6171875, "calib/ece": 0.6316071428571428, "calib/final_conf_rate": 0.4375, "calib/format_rate": 0.328125, "calib/frac_conf_gt_0.9": 0.42857142857142855, "calib/gap": 0.07604298974108459, "calib/mean_conf": 0.8369642857142857, "calib/mu_c": 0.8973913043478261, "calib/mu_w": 0.8213483146067415, "calib/nonempty_final_conf_rate": 0.4375, "calib/nonempty_reasoning_rate": 0.54296875, "calib/nonempty_step_conf_rate": 0.40234375, "calib/pce": 0.6316071428571428, "calib/std_conf": 0.20143609091840964, "calib/step_conf_rate": 0.40234375, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.02734375, "completions/max_length": 2798.0, "completions/max_terminated_length": 2798.0, "completions/mean_length": 443.19921875, "completions/mean_terminated_length": 455.65863037109375, "completions/min_length": 0.0, "completions/min_terminated_length": 2.0, "epoch": 0.020266666666666665, "grad_norm": 0.013113144785165787, "learning_rate": 4.75e-06, "loss": 0.0425, "num_tokens": 6070553.0, "reward": 0.41076862812042236, "reward_std": 0.5509061813354492, "rewards/accuracy_reward_step": 0.10546875, "rewards/final_brier_reward_step": 0.13096562027931213, "rewards/format_reward_step": 0.328125, "rewards/stepwise_brier_reward": 0.22304634749889374, "step": 19 }, { "calib/answer_extract_rate": 0.64453125, "calib/auroc": 0.453448275862069, "calib/avg_num_step_conf": 2.5546875, "calib/ece": 0.6106622516556293, "calib/final_conf_rate": 0.58984375, "calib/format_rate": 0.48046875, "calib/frac_conf_gt_0.9": 0.39072847682119205, "calib/gap": -0.018113300492611062, "calib/mean_conf": 0.8256291390728477, "calib/mu_c": 0.8117142857142856, "calib/mu_w": 0.8298275862068967, "calib/nonempty_final_conf_rate": 0.58984375, "calib/nonempty_reasoning_rate": 0.75390625, "calib/nonempty_step_conf_rate": 0.625, "calib/pce": 0.6022516556291392, "calib/std_conf": 0.2035628876223257, "calib/step_conf_rate": 0.625, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.02734375, "completions/max_length": 3007.0, "completions/max_terminated_length": 3007.0, "completions/mean_length": 402.796875, "completions/mean_terminated_length": 414.1204528808594, "completions/min_length": 0.0, "completions/min_terminated_length": 10.0, "epoch": 0.021333333333333333, "grad_norm": 0.011112475767731667, "learning_rate": 5e-06, "loss": 0.1319, "num_tokens": 6332045.0, "reward": 0.5859308242797852, "reward_std": 0.6181765794754028, "rewards/accuracy_reward_step": 0.13671875, "rewards/final_brier_reward_step": 0.20985156297683716, "rewards/format_reward_step": 0.48046875, "rewards/stepwise_brier_reward": 0.35262158513069153, "step": 20 }, { "calib/answer_extract_rate": 0.76953125, "calib/auroc": 0.5793789808917198, "calib/avg_num_step_conf": 3.1171875, "calib/ece": 0.6150390862944164, "calib/final_conf_rate": 0.76953125, "calib/format_rate": 0.59765625, "calib/frac_conf_gt_0.9": 0.4365482233502538, "calib/gap": 0.07110222929936294, "calib/mean_conf": 0.818084771573604, "calib/mu_c": 0.87475, "calib/mu_w": 0.8036477707006371, "calib/nonempty_final_conf_rate": 0.76953125, "calib/nonempty_reasoning_rate": 0.87109375, "calib/nonempty_step_conf_rate": 0.73828125, "calib/pce": 0.6150390862944164, "calib/std_conf": 0.21389979002838316, "calib/step_conf_rate": 0.73828125, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1837.0, "completions/max_terminated_length": 1837.0, "completions/mean_length": 321.3359375, "completions/mean_terminated_length": 321.3359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.0224, "grad_norm": 0.012988156639039516, "learning_rate": 4.9722222222222224e-06, "loss": 0.0565, "num_tokens": 6570771.0, "reward": 0.7183767557144165, "reward_std": 0.6872435212135315, "rewards/accuracy_reward_step": 0.1640625, "rewards/final_brier_reward_step": 0.26491594314575195, "rewards/format_reward_step": 0.59765625, "rewards/stepwise_brier_reward": 0.4289034307003021, "step": 21 }, { "calib/answer_extract_rate": 0.875, "calib/auroc": 0.5034090909090909, "calib/avg_num_step_conf": 3.34375, "calib/ece": 0.562293023255814, "calib/final_conf_rate": 0.83984375, "calib/format_rate": 0.7421875, "calib/frac_conf_gt_0.9": 0.3627906976744186, "calib/gap": 0.013271590909090913, "calib/mean_conf": 0.791232558139535, "calib/mu_c": 0.8011090909090909, "calib/mu_w": 0.7878375, "calib/nonempty_final_conf_rate": 0.83984375, "calib/nonempty_reasoning_rate": 0.94921875, "calib/nonempty_step_conf_rate": 0.8671875, "calib/pce": 0.5488558139534884, "calib/std_conf": 0.24719120150757437, "calib/step_conf_rate": 0.8671875, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2843.0, "completions/max_terminated_length": 2843.0, "completions/mean_length": 330.20703125, "completions/mean_terminated_length": 330.20703125, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.023466666666666667, "grad_norm": 0.011182010173797607, "learning_rate": 4.944444444444445e-06, "loss": 0.0756, "num_tokens": 6810624.0, "reward": 0.9263468384742737, "reward_std": 0.6675437688827515, "rewards/accuracy_reward_step": 0.22265625, "rewards/final_brier_reward_step": 0.34555596113204956, "rewards/format_reward_step": 0.7421875, "rewards/stepwise_brier_reward": 0.5395188331604004, "step": 22 }, { "calib/answer_extract_rate": 0.90234375, "calib/auroc": 0.47671897584142064, "calib/avg_num_step_conf": 3.5859375, "calib/ece": 0.6022222222222222, "calib/final_conf_rate": 0.87890625, "calib/format_rate": 0.8203125, "calib/frac_conf_gt_0.9": 0.3511111111111111, "calib/gap": -0.04532211439190581, "calib/mean_conf": 0.8034666666666666, "calib/mu_c": 0.7698275862068965, "calib/mu_w": 0.8151497005988023, "calib/nonempty_final_conf_rate": 0.87890625, "calib/nonempty_reasoning_rate": 0.94140625, "calib/nonempty_step_conf_rate": 0.88671875, "calib/pce": 0.5739555555555556, "calib/std_conf": 0.2294311419334544, "calib/step_conf_rate": 0.88671875, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2656.0, "completions/max_terminated_length": 2656.0, "completions/mean_length": 306.859375, "completions/mean_terminated_length": 310.498046875, "completions/min_length": 0.0, "completions/min_terminated_length": 24.0, "epoch": 0.024533333333333334, "grad_norm": 0.008833150379359722, "learning_rate": 4.9166666666666665e-06, "loss": 0.0801, "num_tokens": 7046620.0, "reward": 0.9918270707130432, "reward_std": 0.7378478050231934, "rewards/accuracy_reward_step": 0.23046875, "rewards/final_brier_reward_step": 0.3657972812652588, "rewards/format_reward_step": 0.8203125, "rewards/stepwise_brier_reward": 0.5780737400054932, "step": 23 }, { "calib/answer_extract_rate": 0.921875, "calib/auroc": 0.47305290831416363, "calib/avg_num_step_conf": 4.22265625, "calib/ece": 0.602086956521739, "calib/final_conf_rate": 0.8984375, "calib/format_rate": 0.8515625, "calib/frac_conf_gt_0.9": 0.36086956521739133, "calib/gap": -0.0013013473545843146, "calib/mean_conf": 0.8151304347826088, "calib/mu_c": 0.8141176470588235, "calib/mu_w": 0.8154189944134078, "calib/nonempty_final_conf_rate": 0.8984375, "calib/nonempty_reasoning_rate": 0.984375, "calib/nonempty_step_conf_rate": 0.94140625, "calib/pce": 0.5977391304347826, "calib/std_conf": 0.20431801817750894, "calib/step_conf_rate": 0.94140625, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0078125, "completions/max_length": 2098.0, "completions/max_terminated_length": 2098.0, "completions/mean_length": 284.796875, "completions/mean_terminated_length": 287.03936767578125, "completions/min_length": 0.0, "completions/min_terminated_length": 12.0, "epoch": 0.0256, "grad_norm": 0.009028051979839802, "learning_rate": 4.888888888888889e-06, "loss": 0.0348, "num_tokens": 7277544.0, "reward": 0.9712204933166504, "reward_std": 0.5941590070724487, "rewards/accuracy_reward_step": 0.19921875, "rewards/final_brier_reward_step": 0.36735236644744873, "rewards/format_reward_step": 0.8515625, "rewards/stepwise_brier_reward": 0.6190922260284424, "step": 24 }, { "calib/answer_extract_rate": 0.91796875, "calib/auroc": 0.5543298400441258, "calib/avg_num_step_conf": 4.31640625, "calib/ece": 0.6138034188034188, "calib/final_conf_rate": 0.9140625, "calib/format_rate": 0.83203125, "calib/frac_conf_gt_0.9": 0.4358974358974359, "calib/gap": 0.07906563706563718, "calib/mean_conf": 0.8232051282051281, "calib/mu_c": 0.8857142857142858, "calib/mu_w": 0.8066486486486486, "calib/nonempty_final_conf_rate": 0.9140625, "calib/nonempty_reasoning_rate": 0.96484375, "calib/nonempty_step_conf_rate": 0.9140625, "calib/pce": 0.6138034188034188, "calib/std_conf": 0.21179630750118708, "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": 0.0, "completions/max_length": 2951.0, "completions/max_terminated_length": 2951.0, "completions/mean_length": 311.421875, "completions/mean_terminated_length": 311.421875, "completions/min_length": 32.0, "completions/min_terminated_length": 32.0, "epoch": 0.02666666666666667, "grad_norm": 0.00787280686199665, "learning_rate": 4.861111111111111e-06, "loss": 0.081, "num_tokens": 7513996.0, "reward": 0.9549209475517273, "reward_std": 0.6096559762954712, "rewards/accuracy_reward_step": 0.19921875, "rewards/final_brier_reward_step": 0.35550546646118164, "rewards/format_reward_step": 0.83203125, "rewards/stepwise_brier_reward": 0.6048034429550171, "step": 25 }, { "calib/answer_extract_rate": 0.97265625, "calib/auroc": 0.579321472178615, "calib/avg_num_step_conf": 4.81640625, "calib/ece": 0.5789795918367346, "calib/final_conf_rate": 0.95703125, "calib/format_rate": 0.92578125, "calib/frac_conf_gt_0.9": 0.3836734693877551, "calib/gap": 0.057356532356532375, "calib/mean_conf": 0.8361224489795916, "calib/mu_c": 0.8787301587301588, "calib/mu_w": 0.8213736263736264, "calib/nonempty_final_conf_rate": 0.95703125, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.9765625, "calib/pce": 0.5789795918367346, "calib/std_conf": 0.15964266636652855, "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": 1160.0, "completions/max_terminated_length": 1160.0, "completions/mean_length": 300.9765625, "completions/mean_terminated_length": 303.3464660644531, "completions/min_length": 0.0, "completions/min_terminated_length": 65.0, "epoch": 0.027733333333333332, "grad_norm": 0.007727249059826136, "learning_rate": 4.833333333333333e-06, "loss": 0.0244, "num_tokens": 7749790.0, "reward": 1.107654094696045, "reward_std": 0.6416740417480469, "rewards/accuracy_reward_step": 0.24609375, "rewards/final_brier_reward_step": 0.43978631496429443, "rewards/format_reward_step": 0.92578125, "rewards/stepwise_brier_reward": 0.6627048254013062, "step": 26 }, { "calib/answer_extract_rate": 0.97265625, "calib/auroc": 0.6045490822027135, "calib/avg_num_step_conf": 4.75, "calib/ece": 0.5583471074380165, "calib/final_conf_rate": 0.9453125, "calib/format_rate": 0.91015625, "calib/frac_conf_gt_0.9": 0.33884297520661155, "calib/gap": 0.07668174159794294, "calib/mean_conf": 0.8186776859504132, "calib/mu_c": 0.8753968253968256, "calib/mu_w": 0.7987150837988827, "calib/nonempty_final_conf_rate": 0.9453125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.9765625, "calib/pce": 0.5583471074380165, "calib/std_conf": 0.19274417656898676, "calib/step_conf_rate": 0.9765625, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1008.0, "completions/max_terminated_length": 1008.0, "completions/mean_length": 281.44140625, "completions/mean_terminated_length": 282.54510498046875, "completions/min_length": 0.0, "completions/min_terminated_length": 46.0, "epoch": 0.0288, "grad_norm": 0.0077114952728152275, "learning_rate": 4.805555555555556e-06, "loss": 0.0478, "num_tokens": 7980559.0, "reward": 1.1087486743927002, "reward_std": 0.5898615121841431, "rewards/accuracy_reward_step": 0.25, "rewards/final_brier_reward_step": 0.44046446681022644, "rewards/format_reward_step": 0.91015625, "rewards/stepwise_brier_reward": 0.674217700958252, "step": 27 }, { "calib/answer_extract_rate": 0.96875, "calib/auroc": 0.5565558633425669, "calib/avg_num_step_conf": 5.765625, "calib/ece": 0.5116194331983805, "calib/final_conf_rate": 0.96484375, "calib/format_rate": 0.94140625, "calib/frac_conf_gt_0.9": 0.3562753036437247, "calib/gap": 0.054400584795321505, "calib/mean_conf": 0.8193117408906883, "calib/mu_c": 0.8569736842105262, "calib/mu_w": 0.8025730994152047, "calib/nonempty_final_conf_rate": 0.96484375, "calib/nonempty_reasoning_rate": 0.9921875, "calib/nonempty_step_conf_rate": 0.9765625, "calib/pce": 0.5116194331983805, "calib/std_conf": 0.18630199457329902, "calib/step_conf_rate": 0.9765625, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2152.0, "completions/max_terminated_length": 2152.0, "completions/mean_length": 346.69140625, "completions/mean_terminated_length": 348.0509948730469, "completions/min_length": 0.0, "completions/min_terminated_length": 70.0, "epoch": 0.029866666666666666, "grad_norm": 0.00658813863992691, "learning_rate": 4.777777777777778e-06, "loss": 0.0201, "num_tokens": 8229760.0, "reward": 1.2133698463439941, "reward_std": 0.6396453380584717, "rewards/accuracy_reward_step": 0.296875, "rewards/final_brier_reward_step": 0.48912501335144043, "rewards/format_reward_step": 0.94140625, "rewards/stepwise_brier_reward": 0.7002918720245361, "step": 28 }, { "calib/answer_extract_rate": 0.96875, "calib/auroc": 0.4901020408163266, "calib/avg_num_step_conf": 6.109375, "calib/ece": 0.6014227642276422, "calib/final_conf_rate": 0.9609375, "calib/format_rate": 0.94140625, "calib/frac_conf_gt_0.9": 0.3333333333333333, "calib/gap": 0.006442857142857195, "calib/mean_conf": 0.7916666666666666, "calib/mu_c": 0.7968000000000001, "calib/mu_w": 0.7903571428571429, "calib/nonempty_final_conf_rate": 0.9609375, "calib/nonempty_reasoning_rate": 0.9921875, "calib/nonempty_step_conf_rate": 0.984375, "calib/pce": 0.5949186991869917, "calib/std_conf": 0.22299528474604613, "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": 1547.0, "completions/max_terminated_length": 1547.0, "completions/mean_length": 348.45703125, "completions/mean_terminated_length": 348.45703125, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.030933333333333334, "grad_norm": 0.006569929886609316, "learning_rate": 4.75e-06, "loss": 0.0327, "num_tokens": 8479597.0, "reward": 1.048470139503479, "reward_std": 0.5458044409751892, "rewards/accuracy_reward_step": 0.1953125, "rewards/final_brier_reward_step": 0.4187656342983246, "rewards/format_reward_step": 0.94140625, "rewards/stepwise_brier_reward": 0.7204275131225586, "step": 29 }, { "calib/answer_extract_rate": 0.94921875, "calib/auroc": 0.5764285714285714, "calib/avg_num_step_conf": 5.6875, "calib/ece": 0.5931300813008129, "calib/final_conf_rate": 0.9609375, "calib/format_rate": 0.94140625, "calib/frac_conf_gt_0.9": 0.2764227642276423, "calib/gap": 0.05713469387755121, "calib/mean_conf": 0.7898780487804877, "calib/mu_c": 0.8354, "calib/mu_w": 0.7782653061224488, "calib/nonempty_final_conf_rate": 0.9609375, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.5898780487804878, "calib/std_conf": 0.19886769247829045, "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": 1973.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 338.1171875, "completions/mean_terminated_length": 338.1171875, "completions/min_length": 71.0, "completions/min_terminated_length": 71.0, "epoch": 0.032, "grad_norm": 0.006622521672397852, "learning_rate": 4.722222222222222e-06, "loss": 0.0663, "num_tokens": 8726643.0, "reward": 1.0522184371948242, "reward_std": 0.5504353642463684, "rewards/accuracy_reward_step": 0.1953125, "rewards/final_brier_reward_step": 0.45076286792755127, "rewards/format_reward_step": 0.94140625, "rewards/stepwise_brier_reward": 0.7034236192703247, "step": 30 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.5243969298245613, "calib/avg_num_step_conf": 6.21484375, "calib/ece": 0.5525702811244979, "calib/final_conf_rate": 0.97265625, "calib/format_rate": 0.97265625, "calib/frac_conf_gt_0.9": 0.21686746987951808, "calib/gap": 0.02433662280701754, "calib/mean_conf": 0.7775502008032129, "calib/mu_c": 0.7963157894736842, "calib/mu_w": 0.7719791666666667, "calib/nonempty_final_conf_rate": 0.97265625, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.5506024096385542, "calib/std_conf": 0.19346352377148554, "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": 1677.0, "completions/max_terminated_length": 1677.0, "completions/mean_length": 323.45703125, "completions/mean_terminated_length": 324.7254943847656, "completions/min_length": 0.0, "completions/min_terminated_length": 67.0, "epoch": 0.03306666666666667, "grad_norm": 0.006582073867321014, "learning_rate": 4.694444444444445e-06, "loss": 0.0547, "num_tokens": 8968864.0, "reward": 1.1258108615875244, "reward_std": 0.45009851455688477, "rewards/accuracy_reward_step": 0.22265625, "rewards/final_brier_reward_step": 0.4801519513130188, "rewards/format_reward_step": 0.97265625, "rewards/stepwise_brier_reward": 0.7418413758277893, "step": 31 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.536980198019802, "calib/avg_num_step_conf": 5.7421875, "calib/ece": 0.5244444444444444, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.96484375, "calib/frac_conf_gt_0.9": 0.1388888888888889, "calib/gap": 0.05328712871287122, "calib/mean_conf": 0.7192857142857142, "calib/mu_c": 0.762, "calib/mu_w": 0.7087128712871288, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.9921875, "calib/pce": 0.5226587301587301, "calib/std_conf": 0.23860932595110151, "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": 1846.0, "completions/max_terminated_length": 1846.0, "completions/mean_length": 312.9375, "completions/mean_terminated_length": 312.9375, "completions/min_length": 64.0, "completions/min_terminated_length": 64.0, "epoch": 0.034133333333333335, "grad_norm": 0.006742473226040602, "learning_rate": 4.666666666666667e-06, "loss": 0.052, "num_tokens": 9209184.0, "reward": 1.0940580368041992, "reward_std": 0.4264279007911682, "rewards/accuracy_reward_step": 0.1953125, "rewards/final_brier_reward_step": 0.5114519596099854, "rewards/format_reward_step": 0.96484375, "rewards/stepwise_brier_reward": 0.7632179260253906, "step": 32 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.5147973693565588, "calib/avg_num_step_conf": 6.11328125, "calib/ece": 0.4782142857142858, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.96875, "calib/frac_conf_gt_0.9": 0.12301587301587301, "calib/gap": 0.03162282260931393, "calib/mean_conf": 0.6978968253968254, "calib/mu_c": 0.7222413793103448, "calib/mu_w": 0.6906185567010309, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.47297619047619055, "calib/std_conf": 0.23091592400303593, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 997.0, "completions/max_terminated_length": 997.0, "completions/mean_length": 324.84375, "completions/mean_terminated_length": 326.11767578125, "completions/min_length": 0.0, "completions/min_terminated_length": 80.0, "epoch": 0.0352, "grad_norm": 0.006110044661909342, "learning_rate": 4.638888888888889e-06, "loss": 0.0152, "num_tokens": 9452720.0, "reward": 1.1544158458709717, "reward_std": 0.5174895524978638, "rewards/accuracy_reward_step": 0.2265625, "rewards/final_brier_reward_step": 0.5472074747085571, "rewards/format_reward_step": 0.96875, "rewards/stepwise_brier_reward": 0.7735810875892639, "step": 33 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5147887934773181, "calib/avg_num_step_conf": 5.30859375, "calib/ece": 0.38432, "calib/final_conf_rate": 0.9765625, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.072, "calib/gap": 0.03444878133402729, "calib/mean_conf": 0.61576, "calib/mu_c": 0.6418032786885246, "calib/mu_w": 0.6073544973544973, "calib/nonempty_final_conf_rate": 0.9765625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.37803999999999993, "calib/std_conf": 0.27491893787078403, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1900.0, "completions/max_terminated_length": 1900.0, "completions/mean_length": 277.1484375, "completions/mean_terminated_length": 278.2353210449219, "completions/min_length": 0.0, "completions/min_terminated_length": 61.0, "epoch": 0.03626666666666667, "grad_norm": 0.007039043586701155, "learning_rate": 4.611111111111112e-06, "loss": 0.0301, "num_tokens": 9682286.0, "reward": 1.2069661617279053, "reward_std": 0.5460448265075684, "rewards/accuracy_reward_step": 0.24609375, "rewards/final_brier_reward_step": 0.60005784034729, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.798119306564331, "step": 34 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5720671791443852, "calib/avg_num_step_conf": 6.9765625, "calib/ece": 0.3190836653386454, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.97265625, "calib/frac_conf_gt_0.9": 0.05976095617529881, "calib/gap": 0.06723178475935832, "calib/mean_conf": 0.5294422310756973, "calib/mu_c": 0.57953125, "calib/mu_w": 0.5122994652406417, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.9921875, "calib/pce": 0.29677290836653386, "calib/std_conf": 0.27762741170583366, "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": 1726.0, "completions/max_terminated_length": 1726.0, "completions/mean_length": 354.98046875, "completions/mean_terminated_length": 354.98046875, "completions/min_length": 73.0, "completions/min_terminated_length": 73.0, "epoch": 0.037333333333333336, "grad_norm": 0.005957501009106636, "learning_rate": 4.583333333333333e-06, "loss": 0.0579, "num_tokens": 9935921.0, "reward": 1.238525390625, "reward_std": 0.5458922386169434, "rewards/accuracy_reward_step": 0.25390625, "rewards/final_brier_reward_step": 0.6667132377624512, "rewards/format_reward_step": 0.97265625, "rewards/stepwise_brier_reward": 0.8186385631561279, "step": 35 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.44578194699553925, "calib/avg_num_step_conf": 6.4453125, "calib/ece": 0.28270916334661356, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.03187250996015936, "calib/gap": -0.05176594069797952, "calib/mean_conf": 0.47227091633466134, "calib/mu_c": 0.44174757281553406, "calib/mu_w": 0.4935135135135136, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.17231075697211157, "calib/std_conf": 0.27032668430211226, "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": 1545.0, "completions/max_terminated_length": 1545.0, "completions/mean_length": 307.91796875, "completions/mean_terminated_length": 309.1255187988281, "completions/min_length": 0.0, "completions/min_terminated_length": 51.0, "epoch": 0.0384, "grad_norm": 0.006046306807547808, "learning_rate": 4.555555555555556e-06, "loss": 0.0559, "num_tokens": 10170964.0, "reward": 1.4502986669540405, "reward_std": 0.48770377039909363, "rewards/accuracy_reward_step": 0.40234375, "rewards/final_brier_reward_step": 0.6395988464355469, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.7944082617759705, "step": 36 }, { "calib/answer_extract_rate": 0.953125, "calib/auroc": 0.5523316062176167, "calib/avg_num_step_conf": 6.93359375, "calib/ece": 0.24440329218106996, "calib/final_conf_rate": 0.94921875, "calib/format_rate": 0.9453125, "calib/frac_conf_gt_0.9": 0.02880658436213992, "calib/gap": 0.03785388601036266, "calib/mean_conf": 0.4005349794238683, "calib/mu_c": 0.43060000000000004, "calib/mu_w": 0.3927461139896374, "calib/nonempty_final_conf_rate": 0.94921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.21958847736625514, "calib/std_conf": 0.2761018364391694, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2333.0, "completions/max_terminated_length": 2333.0, "completions/mean_length": 356.546875, "completions/mean_terminated_length": 357.94512939453125, "completions/min_length": 0.0, "completions/min_terminated_length": 51.0, "epoch": 0.039466666666666664, "grad_norm": 0.005678115412592888, "learning_rate": 4.527777777777778e-06, "loss": 0.1154, "num_tokens": 10422840.0, "reward": 1.141843318939209, "reward_std": 0.4752626419067383, "rewards/accuracy_reward_step": 0.1953125, "rewards/final_brier_reward_step": 0.6938046813011169, "rewards/format_reward_step": 0.9453125, "rewards/stepwise_brier_reward": 0.8110687136650085, "step": 37 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.5279794313369631, "calib/avg_num_step_conf": 7.171875, "calib/ece": 0.21664, "calib/final_conf_rate": 0.9765625, "calib/format_rate": 0.96875, "calib/frac_conf_gt_0.9": 0.012, "calib/gap": 0.024842710223835496, "calib/mean_conf": 0.36192, "calib/mu_c": 0.3792105263157895, "calib/mu_w": 0.354367816091954, "calib/nonempty_final_conf_rate": 0.9765625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.13728, "calib/std_conf": 0.2634029491102937, "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": 2085.0, "completions/max_terminated_length": 2085.0, "completions/mean_length": 354.7890625, "completions/mean_terminated_length": 354.7890625, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "epoch": 0.04053333333333333, "grad_norm": 0.005945122335106134, "learning_rate": 4.5e-06, "loss": 0.1174, "num_tokens": 10674058.0, "reward": 1.3120146989822388, "reward_std": 0.4244992136955261, "rewards/accuracy_reward_step": 0.296875, "rewards/final_brier_reward_step": 0.7053617238998413, "rewards/format_reward_step": 0.96875, "rewards/stepwise_brier_reward": 0.8239471316337585, "step": 38 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5830742659758203, "calib/avg_num_step_conf": 6.02734375, "calib/ece": 0.16498814229249015, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.003952569169960474, "calib/gap": 0.04546113989637307, "calib/mean_conf": 0.28932015810276684, "calib/mu_c": 0.324, "calib/mu_w": 0.27853886010362694, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.10857707509881423, "calib/std_conf": 0.24816852513237486, "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": 2070.0, "completions/max_terminated_length": 2070.0, "completions/mean_length": 314.8515625, "completions/mean_terminated_length": 314.8515625, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "epoch": 0.0416, "grad_norm": 0.006720567587763071, "learning_rate": 4.472222222222223e-06, "loss": 0.0294, "num_tokens": 10914252.0, "reward": 1.2488646507263184, "reward_std": 0.35320043563842773, "rewards/accuracy_reward_step": 0.234375, "rewards/final_brier_reward_step": 0.7589088678359985, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8615494966506958, "step": 39 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5632592710997443, "calib/avg_num_step_conf": 6.74609375, "calib/ece": 0.14884920634920634, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.011904761904761904, "calib/gap": 0.04003196930946307, "calib/mean_conf": 0.2679761904761905, "calib/mu_c": 0.29720588235294126, "calib/mu_w": 0.2571739130434782, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.07349206349206348, "calib/std_conf": 0.2342165923488754, "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": 2391.0, "completions/max_terminated_length": 2391.0, "completions/mean_length": 358.2421875, "completions/mean_terminated_length": 358.2421875, "completions/min_length": 68.0, "completions/min_terminated_length": 68.0, "epoch": 0.042666666666666665, "grad_norm": 0.0062150657176971436, "learning_rate": 4.444444444444444e-06, "loss": 0.0905, "num_tokens": 11166226.0, "reward": 1.2930632829666138, "reward_std": 0.47657638788223267, "rewards/accuracy_reward_step": 0.265625, "rewards/final_brier_reward_step": 0.7519511580467224, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.857802152633667, "step": 40 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5062084535306661, "calib/avg_num_step_conf": 6.63671875, "calib/ece": 0.3147233201581027, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.003952569169960474, "calib/gap": 0.0007183619716543688, "calib/mean_conf": 0.20970355731225293, "calib/mu_c": 0.2100840336134454, "calib/mu_w": 0.20936567164179104, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.02703557312252964, "calib/std_conf": 0.2086101453768343, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1012.0, "completions/max_terminated_length": 1012.0, "completions/mean_length": 324.76953125, "completions/mean_terminated_length": 326.04315185546875, "completions/min_length": 0.0, "completions/min_terminated_length": 55.0, "epoch": 0.04373333333333333, "grad_norm": 0.005901300814002752, "learning_rate": 4.416666666666667e-06, "loss": 0.0242, "num_tokens": 11410119.0, "reward": 1.5530972480773926, "reward_std": 0.5297301411628723, "rewards/accuracy_reward_step": 0.46875, "rewards/final_brier_reward_step": 0.6322815418243408, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.791045069694519, "step": 41 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.4984800422944754, "calib/avg_num_step_conf": 6.05078125, "calib/ece": 0.27391304347826084, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.007905138339920948, "calib/gap": -0.0022012952683055365, "calib/mean_conf": 0.17960474308300395, "calib/mu_c": 0.1782474226804124, "calib/mu_w": 0.18044871794871795, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.03505928853754941, "calib/std_conf": 0.2064547388339375, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1814.0, "completions/max_terminated_length": 1814.0, "completions/mean_length": 291.6953125, "completions/mean_terminated_length": 291.6953125, "completions/min_length": 55.0, "completions/min_terminated_length": 55.0, "epoch": 0.0448, "grad_norm": 0.00655122147873044, "learning_rate": 4.388888888888889e-06, "loss": 0.0413, "num_tokens": 11642665.0, "reward": 1.4383958578109741, "reward_std": 0.37608566880226135, "rewards/accuracy_reward_step": 0.37890625, "rewards/final_brier_reward_step": 0.6704492568969727, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8331342339515686, "step": 42 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5327291381668946, "calib/avg_num_step_conf": 5.9453125, "calib/ece": 0.24957031250000003, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.00390625, "calib/gap": -0.008255813953488372, "calib/mean_conf": 0.1522265625, "calib/mu_c": 0.14674418604651163, "calib/mu_w": 0.155, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.0329296875, "calib/std_conf": 0.17660238332036632, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1300.0, "completions/max_terminated_length": 1300.0, "completions/mean_length": 313.61328125, "completions/mean_terminated_length": 314.8431701660156, "completions/min_length": 0.0, "completions/min_terminated_length": 52.0, "epoch": 0.04586666666666667, "grad_norm": 0.006789583247154951, "learning_rate": 4.361111111111112e-06, "loss": -0.0582, "num_tokens": 11881678.0, "reward": 1.3885436058044434, "reward_std": 0.38661620020866394, "rewards/accuracy_reward_step": 0.3359375, "rewards/final_brier_reward_step": 0.7082949876785278, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.8302544355392456, "step": 43 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.6158761100281567, "calib/avg_num_step_conf": 6.94921875, "calib/ece": 0.22214285714285711, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.04369070825211176, "calib/mean_conf": 0.12269841269841271, "calib/mu_c": 0.15234567901234566, "calib/mu_w": 0.10865497076023391, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.011706349206349205, "calib/std_conf": 0.14737067345406912, "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": 2416.0, "completions/max_terminated_length": 2416.0, "completions/mean_length": 372.20703125, "completions/mean_terminated_length": 372.20703125, "completions/min_length": 72.0, "completions/min_terminated_length": 72.0, "epoch": 0.046933333333333334, "grad_norm": 0.005960023496299982, "learning_rate": 4.333333333333334e-06, "loss": 0.083, "num_tokens": 12136787.0, "reward": 1.3599913120269775, "reward_std": 0.35860058665275574, "rewards/accuracy_reward_step": 0.31640625, "rewards/final_brier_reward_step": 0.7243582010269165, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.8562322854995728, "step": 44 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5408417508417508, "calib/avg_num_step_conf": 5.8984375, "calib/ece": 0.2823117647058823, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.00392156862745098, "calib/gap": 0.006976767676767678, "calib/mean_conf": 0.09604117647058824, "calib/mu_c": 0.10055555555555556, "calib/mu_w": 0.09357878787878789, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.012705882352941176, "calib/std_conf": 0.12731624694386345, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1931.0, "completions/max_terminated_length": 1931.0, "completions/mean_length": 307.4140625, "completions/mean_terminated_length": 307.4140625, "completions/min_length": 56.0, "completions/min_terminated_length": 56.0, "epoch": 0.048, "grad_norm": 0.005965207703411579, "learning_rate": 4.305555555555556e-06, "loss": 0.0429, "num_tokens": 12374037.0, "reward": 1.3967806100845337, "reward_std": 0.42153504490852356, "rewards/accuracy_reward_step": 0.3515625, "rewards/final_brier_reward_step": 0.6844413876533508, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8167434930801392, "step": 45 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.5096754887618034, "calib/avg_num_step_conf": 7.05859375, "calib/ece": 0.3476305220883534, "calib/final_conf_rate": 0.97265625, "calib/format_rate": 0.97265625, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.0011411091900518783, "calib/mean_conf": 0.07855421686746988, "calib/mu_c": 0.0792233009708738, "calib/mu_w": 0.07808219178082192, "calib/nonempty_final_conf_rate": 0.97265625, "calib/nonempty_reasoning_rate": 0.9921875, "calib/nonempty_step_conf_rate": 0.9921875, "calib/pce": 0.006265060240963856, "calib/std_conf": 0.09727779783741791, "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": 2238.0, "completions/max_terminated_length": 2238.0, "completions/mean_length": 388.45703125, "completions/mean_terminated_length": 389.98040771484375, "completions/min_length": 0.0, "completions/min_terminated_length": 90.0, "epoch": 0.04906666666666667, "grad_norm": 0.006083706859499216, "learning_rate": 4.277777777777778e-06, "loss": 0.0236, "num_tokens": 12631754.0, "reward": 1.4434081315994263, "reward_std": 0.4992774724960327, "rewards/accuracy_reward_step": 0.40234375, "rewards/final_brier_reward_step": 0.6188562512397766, "rewards/format_reward_step": 0.97265625, "rewards/stepwise_brier_reward": 0.7954011559486389, "step": 46 }, { "calib/answer_extract_rate": 0.97265625, "calib/auroc": 0.4819284884471408, "calib/avg_num_step_conf": 7.8359375, "calib/ece": 0.4416064257028112, "calib/final_conf_rate": 0.97265625, "calib/format_rate": 0.97265625, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": -0.009050600232347994, "calib/mean_conf": 0.07726907630522088, "calib/mu_c": 0.07283464566929135, "calib/mu_w": 0.08188524590163934, "calib/nonempty_final_conf_rate": 0.97265625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.004417670682730922, "calib/std_conf": 0.07601899392454098, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2512.0, "completions/max_terminated_length": 2512.0, "completions/mean_length": 411.0390625, "completions/mean_terminated_length": 412.6510009765625, "completions/min_length": 0.0, "completions/min_terminated_length": 85.0, "epoch": 0.050133333333333335, "grad_norm": 0.005414419341832399, "learning_rate": 4.25e-06, "loss": 0.1349, "num_tokens": 12896460.0, "reward": 1.5517261028289795, "reward_std": 0.4409915804862976, "rewards/accuracy_reward_step": 0.49609375, "rewards/final_brier_reward_step": 0.5374000072479248, "rewards/format_reward_step": 0.97265625, "rewards/stepwise_brier_reward": 0.747629702091217, "step": 47 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5095758757573647, "calib/avg_num_step_conf": 5.66796875, "calib/ece": 0.259375, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.009056340970819685, "calib/mean_conf": 0.06484375, "calib/mu_c": 0.07096385542168673, "calib/mu_w": 0.06190751445086705, "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": 0.06549671240556659, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 961.0, "completions/max_terminated_length": 961.0, "completions/mean_length": 315.05078125, "completions/mean_terminated_length": 316.2862854003906, "completions/min_length": 0.0, "completions/min_terminated_length": 54.0, "epoch": 0.0512, "grad_norm": 0.006967704743146896, "learning_rate": 4.222222222222223e-06, "loss": -0.0318, "num_tokens": 13134305.0, "reward": 1.3742893934249878, "reward_std": 0.36246737837791443, "rewards/accuracy_reward_step": 0.32421875, "rewards/final_brier_reward_step": 0.7133023738861084, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.8385426998138428, "step": 48 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.5701233148041659, "calib/avg_num_step_conf": 6.3359375, "calib/ece": 0.37476190476190474, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.00975273145485911, "calib/mean_conf": 0.06571428571428573, "calib/mu_c": 0.07117117117117117, "calib/mu_w": 0.06141843971631206, "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": 0.05889765903975142, "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": 1896.0, "completions/max_terminated_length": 1896.0, "completions/mean_length": 337.0703125, "completions/mean_terminated_length": 338.3921813964844, "completions/min_length": 0.0, "completions/min_terminated_length": 64.0, "epoch": 0.05226666666666667, "grad_norm": 0.006038546562194824, "learning_rate": 4.194444444444445e-06, "loss": 0.0283, "num_tokens": 13378635.0, "reward": 1.4912843704223633, "reward_std": 0.320556640625, "rewards/accuracy_reward_step": 0.43359375, "rewards/final_brier_reward_step": 0.6009379029273987, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.8016999959945679, "step": 49 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5505583126550868, "calib/avg_num_step_conf": 7.08203125, "calib/ece": 0.4545669291338583, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.00577419354838709, "calib/mean_conf": 0.06118110236220472, "calib/mu_c": 0.064, "calib/mu_w": 0.05822580645161291, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.001968503937007874, "calib/std_conf": 0.05888884729549961, "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": 2281.0, "completions/max_terminated_length": 2281.0, "completions/mean_length": 404.8828125, "completions/mean_terminated_length": 406.4706115722656, "completions/min_length": 0.0, "completions/min_terminated_length": 97.0, "epoch": 0.05333333333333334, "grad_norm": 0.005830938927829266, "learning_rate": 4.166666666666667e-06, "loss": 0.0055, "num_tokens": 13641149.0, "reward": 1.5887231826782227, "reward_std": 0.3642390966415405, "rewards/accuracy_reward_step": 0.5078125, "rewards/final_brier_reward_step": 0.5422203540802002, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.7814227342605591, "step": 50 }, { "calib/answer_extract_rate": 0.96875, "calib/auroc": 0.5659294056240621, "calib/avg_num_step_conf": 7.50390625, "calib/ece": 0.4095564516129032, "calib/final_conf_rate": 0.96875, "calib/format_rate": 0.96875, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.011538461538461567, "calib/mean_conf": 0.06544354838709678, "calib/mu_c": 0.07153846153846156, "calib/mu_w": 0.06, "calib/nonempty_final_conf_rate": 0.96875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.0016129032258064516, "calib/std_conf": 0.07334604266424824, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2424.0, "completions/max_terminated_length": 2424.0, "completions/mean_length": 434.2109375, "completions/mean_terminated_length": 434.2109375, "completions/min_length": 61.0, "completions/min_terminated_length": 61.0, "epoch": 0.0544, "grad_norm": 0.005800957791507244, "learning_rate": 4.138888888888889e-06, "loss": 0.0709, "num_tokens": 13915107.0, "reward": 1.5064377784729004, "reward_std": 0.37380897998809814, "rewards/accuracy_reward_step": 0.45703125, "rewards/final_brier_reward_step": 0.5677487850189209, "rewards/format_reward_step": 0.96875, "rewards/stepwise_brier_reward": 0.7783145308494568, "step": 51 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.6226774193548387, "calib/avg_num_step_conf": 6.171875, "calib/ece": 0.5361176470588236, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.02319677419354839, "calib/mean_conf": 0.07800000000000001, "calib/mu_c": 0.08709677419354839, "calib/mu_w": 0.0639, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.0031372549019607846, "calib/std_conf": 0.08362568326752723, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1631.0, "completions/max_terminated_length": 1631.0, "completions/mean_length": 366.6640625, "completions/mean_terminated_length": 366.6640625, "completions/min_length": 110.0, "completions/min_terminated_length": 110.0, "epoch": 0.055466666666666664, "grad_norm": 0.006987405009567738, "learning_rate": 4.111111111111111e-06, "loss": -0.004, "num_tokens": 14170429.0, "reward": 1.7260549068450928, "reward_std": 0.399250864982605, "rewards/accuracy_reward_step": 0.60546875, "rewards/final_brier_reward_step": 0.4830675721168518, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.7961521744728088, "step": 52 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5364954380702412, "calib/avg_num_step_conf": 7.5078125, "calib/ece": 0.40924901185770746, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.025980502437195338, "calib/mean_conf": 0.09351778656126483, "calib/mu_c": 0.10645669291338584, "calib/mu_w": 0.0804761904761905, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.000395256916996046, "calib/std_conf": 0.10621346932963562, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 1566.0, "completions/max_terminated_length": 1566.0, "completions/mean_length": 424.4140625, "completions/mean_terminated_length": 424.4140625, "completions/min_length": 84.0, "completions/min_terminated_length": 84.0, "epoch": 0.05653333333333333, "grad_norm": 0.006259525194764137, "learning_rate": 4.083333333333334e-06, "loss": -0.0073, "num_tokens": 14438407.0, "reward": 1.576069712638855, "reward_std": 0.31807127594947815, "rewards/accuracy_reward_step": 0.5, "rewards/final_brier_reward_step": 0.5780203342437744, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.7496961355209351, "step": 53 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5786217399581192, "calib/avg_num_step_conf": 6.4296875, "calib/ece": 0.5032421875, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.015410241766609586, "calib/mean_conf": 0.10105468750000002, "calib/mu_c": 0.10725490196078434, "calib/mu_w": 0.09184466019417475, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.0033203125, "calib/std_conf": 0.09690554426490439, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1154.0, "completions/max_terminated_length": 1154.0, "completions/mean_length": 359.3828125, "completions/mean_terminated_length": 360.79217529296875, "completions/min_length": 0.0, "completions/min_terminated_length": 53.0, "epoch": 0.0576, "grad_norm": 0.007786636706441641, "learning_rate": 4.055555555555556e-06, "loss": 0.0096, "num_tokens": 14690145.0, "reward": 1.713097333908081, "reward_std": 0.2552579641342163, "rewards/accuracy_reward_step": 0.59765625, "rewards/final_brier_reward_step": 0.510944128036499, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.7555073499679565, "step": 54 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5805760383185227, "calib/avg_num_step_conf": 7.52734375, "calib/ece": 0.3969047619047618, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.013057288712421997, "calib/mean_conf": 0.12928571428571428, "calib/mu_c": 0.13565891472868216, "calib/mu_w": 0.12260162601626017, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.007142857142857142, "calib/std_conf": 0.11509608526590909, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2213.0, "completions/max_terminated_length": 2213.0, "completions/mean_length": 428.51953125, "completions/mean_terminated_length": 430.20001220703125, "completions/min_length": 0.0, "completions/min_terminated_length": 84.0, "epoch": 0.058666666666666666, "grad_norm": 0.006805142387747765, "learning_rate": 4.027777777777779e-06, "loss": 0.0398, "num_tokens": 14961174.0, "reward": 1.5985276699066162, "reward_std": 0.3979921340942383, "rewards/accuracy_reward_step": 0.50390625, "rewards/final_brier_reward_step": 0.5876937508583069, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8142296671867371, "step": 55 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5168293610547666, "calib/avg_num_step_conf": 8.125, "calib/ece": 0.3303571428571428, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.0009584178498985574, "calib/mean_conf": 0.1625, "calib/mu_c": 0.16301724137931034, "calib/mu_w": 0.16205882352941178, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.016269841269841268, "calib/std_conf": 0.13107196920436923, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1714.0, "completions/max_terminated_length": 1714.0, "completions/mean_length": 448.4296875, "completions/mean_terminated_length": 450.1882629394531, "completions/min_length": 0.0, "completions/min_terminated_length": 127.0, "epoch": 0.05973333333333333, "grad_norm": 0.006386685650795698, "learning_rate": 4.000000000000001e-06, "loss": 0.0164, "num_tokens": 15236316.0, "reward": 1.5380668640136719, "reward_std": 0.4855208992958069, "rewards/accuracy_reward_step": 0.453125, "rewards/final_brier_reward_step": 0.6360793113708496, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8286879062652588, "step": 56 }, { "calib/answer_extract_rate": 0.96875, "calib/auroc": 0.48295030845366416, "calib/avg_num_step_conf": 8.34375, "calib/ece": 0.44504032258064513, "calib/final_conf_rate": 0.96875, "calib/format_rate": 0.96875, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": -0.02159785777235443, "calib/mean_conf": 0.18318548387096775, "calib/mu_c": 0.17456375838926172, "calib/mu_w": 0.19616161616161615, "calib/nonempty_final_conf_rate": 0.96875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.013709677419354839, "calib/std_conf": 0.13799812217840454, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2394.0, "completions/max_terminated_length": 2394.0, "completions/mean_length": 484.109375, "completions/mean_terminated_length": 484.109375, "completions/min_length": 90.0, "completions/min_terminated_length": 90.0, "epoch": 0.0608, "grad_norm": 0.005632797256112099, "learning_rate": 3.972222222222223e-06, "loss": 0.0536, "num_tokens": 15520544.0, "reward": 1.6914198398590088, "reward_std": 0.4557984471321106, "rewards/accuracy_reward_step": 0.58203125, "rewards/final_brier_reward_step": 0.5389652252197266, "rewards/format_reward_step": 0.96875, "rewards/stepwise_brier_reward": 0.7970266342163086, "step": 57 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.5651055069480186, "calib/avg_num_step_conf": 9.4140625, "calib/ece": 0.26835999999999993, "calib/final_conf_rate": 0.9765625, "calib/format_rate": 0.97265625, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.034543232115285594, "calib/mean_conf": 0.23364000000000001, "calib/mu_c": 0.2521551724137931, "calib/mu_w": 0.2176119402985075, "calib/nonempty_final_conf_rate": 0.9765625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.01900000000000001, "calib/std_conf": 0.15918024500546543, "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": 2013.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 530.58984375, "completions/mean_terminated_length": 534.7677001953125, "completions/min_length": 0.0, "completions/min_terminated_length": 165.0, "epoch": 0.06186666666666667, "grad_norm": 0.005325364414602518, "learning_rate": 3.944444444444445e-06, "loss": -0.0027, "num_tokens": 15816199.0, "reward": 1.5344479084014893, "reward_std": 0.5253565907478333, "rewards/accuracy_reward_step": 0.453125, "rewards/final_brier_reward_step": 0.6704726219177246, "rewards/format_reward_step": 0.97265625, "rewards/stepwise_brier_reward": 0.8032563328742981, "step": 58 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.47337240814268605, "calib/avg_num_step_conf": 8.90625, "calib/ece": 0.29948412698412696, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": -0.019045188126299878, "calib/mean_conf": 0.24138888888888888, "calib/mu_c": 0.23209302325581396, "calib/mu_w": 0.25113821138211384, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.014484126984126981, "calib/std_conf": 0.15030938845426653, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2480.0, "completions/max_terminated_length": 2480.0, "completions/mean_length": 524.640625, "completions/mean_terminated_length": 526.6980590820312, "completions/min_length": 0.0, "completions/min_terminated_length": 72.0, "epoch": 0.06293333333333333, "grad_norm": 0.005380685441195965, "learning_rate": 3.916666666666667e-06, "loss": 0.0589, "num_tokens": 16110259.0, "reward": 1.6063846349716187, "reward_std": 0.45577263832092285, "rewards/accuracy_reward_step": 0.50390625, "rewards/final_brier_reward_step": 0.6347769498825073, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.7985740900039673, "step": 59 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.4611180124223602, "calib/avg_num_step_conf": 7.68359375, "calib/ece": 0.27262745098039215, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": -0.02071118012422357, "calib/mean_conf": 0.24658823529411766, "calib/mu_c": 0.23521739130434782, "calib/mu_w": 0.2559285714285714, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.03411764705882353, "calib/std_conf": 0.14675218226263825, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1306.0, "completions/max_terminated_length": 1306.0, "completions/mean_length": 447.96875, "completions/mean_terminated_length": 449.72552490234375, "completions/min_length": 0.0, "completions/min_terminated_length": 53.0, "epoch": 0.064, "grad_norm": 0.0065772817470133305, "learning_rate": 3.88888888888889e-06, "loss": 0.0115, "num_tokens": 16387299.0, "reward": 1.5557947158813477, "reward_std": 0.4793645143508911, "rewards/accuracy_reward_step": 0.44921875, "rewards/final_brier_reward_step": 0.6761828660964966, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.859495997428894, "step": 60 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.4977733174433921, "calib/avg_num_step_conf": 7.67578125, "calib/ece": 0.334296875, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.003058395534090208, "calib/mean_conf": 0.26140625, "calib/mu_c": 0.2626845637583893, "calib/mu_w": 0.25962616822429907, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.006835937499999998, "calib/std_conf": 0.1583100911532095, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/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": 432.484375, "completions/mean_terminated_length": 434.180419921875, "completions/min_length": 0.0, "completions/min_terminated_length": 144.0, "epoch": 0.06506666666666666, "grad_norm": 0.006080757826566696, "learning_rate": 3.861111111111112e-06, "loss": 0.0208, "num_tokens": 16655583.0, "reward": 1.736926555633545, "reward_std": 0.30407387018203735, "rewards/accuracy_reward_step": 0.58203125, "rewards/final_brier_reward_step": 0.6303547024726868, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.8251640200614929, "step": 61 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.4068636982306047, "calib/avg_num_step_conf": 9.90625, "calib/ece": 0.27116, "calib/final_conf_rate": 0.9765625, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": -0.0669460107589604, "calib/mean_conf": 0.33524, "calib/mu_c": 0.298018018018018, "calib/mu_w": 0.3649640287769784, "calib/nonempty_final_conf_rate": 0.9765625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.0812, "calib/std_conf": 0.19754326715937448, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2442.0, "completions/max_terminated_length": 2442.0, "completions/mean_length": 582.06640625, "completions/mean_terminated_length": 582.06640625, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.06613333333333334, "grad_norm": 0.005049685016274452, "learning_rate": 3.833333333333334e-06, "loss": 0.0473, "num_tokens": 16965176.0, "reward": 1.5019876956939697, "reward_std": 0.5126863718032837, "rewards/accuracy_reward_step": 0.43359375, "rewards/final_brier_reward_step": 0.6535457372665405, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.7997174263000488, "step": 62 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5755354773488343, "calib/avg_num_step_conf": 9.0625, "calib/ece": 0.24166666666666664, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.003968253968253968, "calib/gap": 0.042370632463511804, "calib/mean_conf": 0.31269841269841275, "calib/mu_c": 0.33270676691729334, "calib/mu_w": 0.29033613445378154, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.013293650793650792, "calib/std_conf": 0.17678303758684782, "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": 2434.0, "completions/max_terminated_length": 2434.0, "completions/mean_length": 583.17578125, "completions/mean_terminated_length": 583.17578125, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "epoch": 0.0672, "grad_norm": 0.0048263357020914555, "learning_rate": 3.8055555555555556e-06, "loss": 0.0185, "num_tokens": 17276613.0, "reward": 1.6461408138275146, "reward_std": 0.42583319544792175, "rewards/accuracy_reward_step": 0.51953125, "rewards/final_brier_reward_step": 0.6827882528305054, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.8236498832702637, "step": 63 }, { "calib/answer_extract_rate": 0.953125, "calib/auroc": 0.5158941755537326, "calib/avg_num_step_conf": 9.23046875, "calib/ece": 0.24868852459016394, "calib/final_conf_rate": 0.953125, "calib/format_rate": 0.9453125, "calib/frac_conf_gt_0.9": 0.0, "calib/gap": 0.006390484003281516, "calib/mean_conf": 0.3641803278688524, "calib/mu_c": 0.3669565217391305, "calib/mu_w": 0.360566037735849, "calib/nonempty_final_conf_rate": 0.953125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.023647540983606555, "calib/std_conf": 0.19462801916095085, "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": 2422.0, "completions/max_terminated_length": 2422.0, "completions/mean_length": 542.203125, "completions/mean_terminated_length": 546.472412109375, "completions/min_length": 0.0, "completions/min_terminated_length": 146.0, "epoch": 0.06826666666666667, "grad_norm": 0.005591726396232843, "learning_rate": 3.777777777777778e-06, "loss": 0.0618, "num_tokens": 17572697.0, "reward": 1.6354286670684814, "reward_std": 0.4554384648799896, "rewards/accuracy_reward_step": 0.5390625, "rewards/final_brier_reward_step": 0.643588662147522, "rewards/format_reward_step": 0.9453125, "rewards/stepwise_brier_reward": 0.7731260061264038, "step": 64 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5711291836096339, "calib/avg_num_step_conf": 8.07421875, "calib/ece": 0.21791338582677172, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.007874015748031496, "calib/gap": 0.038635595871129136, "calib/mean_conf": 0.37279527559055115, "calib/mu_c": 0.3902877697841727, "calib/mu_w": 0.35165217391304354, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.9921875, "calib/pce": 0.02173228346456693, "calib/std_conf": 0.18219263140508696, "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": 1416.0, "completions/max_terminated_length": 1416.0, "completions/mean_length": 448.7421875, "completions/mean_terminated_length": 450.5019836425781, "completions/min_length": 0.0, "completions/min_terminated_length": 140.0, "epoch": 0.06933333333333333, "grad_norm": 0.005525464192032814, "learning_rate": 3.7500000000000005e-06, "loss": -0.0021, "num_tokens": 17846103.0, "reward": 1.6987894773483276, "reward_std": 0.3502931296825409, "rewards/accuracy_reward_step": 0.54296875, "rewards/final_brier_reward_step": 0.7014793157577515, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.859303891658783, "step": 65 }, { "calib/answer_extract_rate": 0.9453125, "calib/auroc": 0.40809282514503925, "calib/avg_num_step_conf": 9.9140625, "calib/ece": 0.2718181818181818, "calib/final_conf_rate": 0.9453125, "calib/format_rate": 0.9453125, "calib/frac_conf_gt_0.9": 0.02066115702479339, "calib/gap": -0.06656532275831673, "calib/mean_conf": 0.41826446280991736, "calib/mu_c": 0.3761797752808989, "calib/mu_w": 0.44274509803921563, "calib/nonempty_final_conf_rate": 0.9453125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.16115702479338848, "calib/std_conf": 0.21605278543096829, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2712.0, "completions/max_terminated_length": 2712.0, "completions/mean_length": 637.73828125, "completions/mean_terminated_length": 637.73828125, "completions/min_length": 119.0, "completions/min_terminated_length": 119.0, "epoch": 0.0704, "grad_norm": 0.004792076535522938, "learning_rate": 3.7222222222222225e-06, "loss": 0.0874, "num_tokens": 18169220.0, "reward": 1.350499153137207, "reward_std": 0.47670456767082214, "rewards/accuracy_reward_step": 0.34765625, "rewards/final_brier_reward_step": 0.6497148275375366, "rewards/format_reward_step": 0.9453125, "rewards/stepwise_brier_reward": 0.7757192850112915, "step": 66 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5702044123145955, "calib/avg_num_step_conf": 8.5546875, "calib/ece": 0.18101960784313723, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.0196078431372549, "calib/gap": 0.04760189455315972, "calib/mean_conf": 0.4072156862745098, "calib/mu_c": 0.4283098591549296, "calib/mu_w": 0.38070796460176987, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.015686274509803914, "calib/std_conf": 0.1875259264402502, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/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": 537.1171875, "completions/mean_terminated_length": 539.2235717773438, "completions/min_length": 0.0, "completions/min_terminated_length": 166.0, "epoch": 0.07146666666666666, "grad_norm": 0.005597411654889584, "learning_rate": 3.694444444444445e-06, "loss": 0.0607, "num_tokens": 18465234.0, "reward": 1.718228816986084, "reward_std": 0.3464631140232086, "rewards/accuracy_reward_step": 0.5546875, "rewards/final_brier_reward_step": 0.7163569927215576, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8362460136413574, "step": 67 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.610280546327058, "calib/avg_num_step_conf": 9.09375, "calib/ece": 0.14160000000000003, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.0196078431372549, "calib/gap": 0.08034865263935037, "calib/mean_conf": 0.43565490196078427, "calib/mu_c": 0.47535658914728685, "calib/mu_w": 0.3950079365079365, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.03568627450980392, "calib/std_conf": 0.21938030039513798, "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": 1995.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 534.06640625, "completions/mean_terminated_length": 534.06640625, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 0.07253333333333334, "grad_norm": 0.005593921989202499, "learning_rate": 3.6666666666666666e-06, "loss": 0.0188, "num_tokens": 18759547.0, "reward": 1.6340742111206055, "reward_std": 0.39258357882499695, "rewards/accuracy_reward_step": 0.50390625, "rewards/final_brier_reward_step": 0.7227994203567505, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8213100433349609, "step": 68 }, { "calib/answer_extract_rate": 0.96484375, "calib/auroc": 0.5605643044619423, "calib/avg_num_step_conf": 8.53515625, "calib/ece": 0.1482186234817814, "calib/final_conf_rate": 0.96484375, "calib/format_rate": 0.96484375, "calib/frac_conf_gt_0.9": 0.008097165991902834, "calib/gap": 0.04208136482939634, "calib/mean_conf": 0.45469635627530364, "calib/mu_c": 0.47633333333333333, "calib/mu_w": 0.434251968503937, "calib/nonempty_final_conf_rate": 0.96484375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.058542510121457475, "calib/std_conf": 0.20424990871330967, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2527.0, "completions/max_terminated_length": 2527.0, "completions/mean_length": 584.16015625, "completions/mean_terminated_length": 586.4509887695312, "completions/min_length": 0.0, "completions/min_terminated_length": 146.0, "epoch": 0.0736, "grad_norm": 0.005437064450234175, "learning_rate": 3.638888888888889e-06, "loss": 0.0705, "num_tokens": 19067092.0, "reward": 1.5660204887390137, "reward_std": 0.5032174587249756, "rewards/accuracy_reward_step": 0.46875, "rewards/final_brier_reward_step": 0.7029246091842651, "rewards/format_reward_step": 0.96484375, "rewards/stepwise_brier_reward": 0.8189697265625, "step": 69 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.5299783135604031, "calib/avg_num_step_conf": 9.24609375, "calib/ece": 0.17456374501992034, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.05179282868525897, "calib/gap": 0.034645936981757997, "calib/mean_conf": 0.48655179282868527, "calib/mu_c": 0.5027014925373136, "calib/mu_w": 0.46805555555555556, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.06362549800796813, "calib/std_conf": 0.23265488202539808, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2367.0, "completions/max_terminated_length": 2367.0, "completions/mean_length": 598.08984375, "completions/mean_terminated_length": 598.08984375, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.07466666666666667, "grad_norm": 0.005056749098002911, "learning_rate": 3.6111111111111115e-06, "loss": 0.0727, "num_tokens": 19380699.0, "reward": 1.6541014909744263, "reward_std": 0.33991795778274536, "rewards/accuracy_reward_step": 0.5234375, "rewards/final_brier_reward_step": 0.6981168389320374, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.8167266249656677, "step": 70 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.5461731251204935, "calib/avg_num_step_conf": 8.4296875, "calib/ece": 0.15340800000000002, "calib/final_conf_rate": 0.9765625, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.016, "calib/gap": 0.041629458261037255, "calib/mean_conf": 0.4996480000000001, "calib/mu_c": 0.5217948717948718, "calib/mu_w": 0.4801654135338346, "calib/nonempty_final_conf_rate": 0.9765625, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.092528, "calib/std_conf": 0.2044007145192991, "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": 2383.0, "completions/max_terminated_length": 2383.0, "completions/mean_length": 547.3359375, "completions/mean_terminated_length": 547.3359375, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "epoch": 0.07573333333333333, "grad_norm": 0.00498660933226347, "learning_rate": 3.5833333333333335e-06, "loss": 0.04, "num_tokens": 19678729.0, "reward": 1.5554978847503662, "reward_std": 0.4403902590274811, "rewards/accuracy_reward_step": 0.45703125, "rewards/final_brier_reward_step": 0.7118868827819824, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.814792275428772, "step": 71 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5720363017583665, "calib/avg_num_step_conf": 9.2265625, "calib/ece": 0.1480079365079365, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.04365079365079365, "calib/gap": 0.04496691245982243, "calib/mean_conf": 0.5647698412698413, "calib/mu_c": 0.587788617886179, "calib/mu_w": 0.5428217054263565, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.11234126984126983, "calib/std_conf": 0.21618435925313628, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2310.0, "completions/max_terminated_length": 2310.0, "completions/mean_length": 544.59375, "completions/mean_terminated_length": 544.59375, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.0768, "grad_norm": 0.005361003801226616, "learning_rate": 3.555555555555556e-06, "loss": 0.0756, "num_tokens": 19976057.0, "reward": 1.5959906578063965, "reward_std": 0.40934091806411743, "rewards/accuracy_reward_step": 0.48046875, "rewards/final_brier_reward_step": 0.7087478041648865, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8236521482467651, "step": 72 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5395759042282221, "calib/avg_num_step_conf": 8.3203125, "calib/ece": 0.1550117647058823, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.054901960784313725, "calib/gap": 0.03593052725420287, "calib/mean_conf": 0.566478431372549, "calib/mu_c": 0.5811324503311259, "calib/mu_w": 0.5452019230769231, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.06466666666666668, "calib/std_conf": 0.2149467387870457, "calib/step_conf_rate": 0.99609375, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1253.0, "completions/max_terminated_length": 1253.0, "completions/mean_length": 518.22265625, "completions/mean_terminated_length": 520.2549438476562, "completions/min_length": 0.0, "completions/min_terminated_length": 199.0, "epoch": 0.07786666666666667, "grad_norm": 0.004839586094021797, "learning_rate": 3.5277777777777784e-06, "loss": 0.0176, "num_tokens": 20269258.0, "reward": 1.7623991966247559, "reward_std": 0.44041845202445984, "rewards/accuracy_reward_step": 0.58984375, "rewards/final_brier_reward_step": 0.723209023475647, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.8029502034187317, "step": 73 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.4190641893991176, "calib/avg_num_step_conf": 8.41015625, "calib/ece": 0.2571003937007874, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.051181102362204724, "calib/gap": -0.05897231715652773, "calib/mean_conf": 0.5515610236220472, "calib/mu_c": 0.5206818181818181, "calib/mu_w": 0.5796541353383459, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.16614173228346454, "calib/std_conf": 0.2321604395699543, "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": 2387.0, "completions/max_terminated_length": 2387.0, "completions/mean_length": 521.92578125, "completions/mean_terminated_length": 523.9725952148438, "completions/min_length": 0.0, "completions/min_terminated_length": 152.0, "epoch": 0.07893333333333333, "grad_norm": 0.00565366493538022, "learning_rate": 3.5e-06, "loss": -0.0023, "num_tokens": 20560303.0, "reward": 1.5749061107635498, "reward_std": 0.38803696632385254, "rewards/accuracy_reward_step": 0.47265625, "rewards/final_brier_reward_step": 0.653693437576294, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8334314823150635, "step": 74 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.5914237709109503, "calib/avg_num_step_conf": 7.8046875, "calib/ece": 0.14469799999999997, "calib/final_conf_rate": 0.9765625, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.064, "calib/gap": 0.06204437869822488, "calib/mean_conf": 0.601942, "calib/mu_c": 0.6220443786982248, "calib/mu_w": 0.5599999999999999, "calib/nonempty_final_conf_rate": 0.9765625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.03531999999999999, "calib/std_conf": 0.19863719096886162, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2214.0, "completions/max_terminated_length": 2214.0, "completions/mean_length": 507.3125, "completions/mean_terminated_length": 511.3070983886719, "completions/min_length": 0.0, "completions/min_terminated_length": 139.0, "epoch": 0.08, "grad_norm": 0.004746343940496445, "learning_rate": 3.4722222222222224e-06, "loss": -0.0295, "num_tokens": 20848431.0, "reward": 1.8710606098175049, "reward_std": 0.42100781202316284, "rewards/accuracy_reward_step": 0.66015625, "rewards/final_brier_reward_step": 0.7453252673149109, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.8248549699783325, "step": 75 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5313643821784677, "calib/avg_num_step_conf": 7.84765625, "calib/ece": 0.15290711462450593, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.05138339920948617, "calib/gap": 0.028615044806260315, "calib/mean_conf": 0.5913616600790512, "calib/mu_c": 0.6042553956834532, "calib/mu_w": 0.5756403508771929, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.0974308300395257, "calib/std_conf": 0.20839986228689478, "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": 2344.0, "completions/max_terminated_length": 2344.0, "completions/mean_length": 551.87109375, "completions/mean_terminated_length": 551.87109375, "completions/min_length": 129.0, "completions/min_terminated_length": 129.0, "epoch": 0.08106666666666666, "grad_norm": 0.0046466956846416, "learning_rate": 3.444444444444445e-06, "loss": 0.0344, "num_tokens": 21146270.0, "reward": 1.6941250562667847, "reward_std": 0.4450863301753998, "rewards/accuracy_reward_step": 0.54296875, "rewards/final_brier_reward_step": 0.7129640579223633, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8291610479354858, "step": 76 }, { "calib/answer_extract_rate": 0.96484375, "calib/auroc": 0.42035460992907797, "calib/avg_num_step_conf": 7.5390625, "calib/ece": 0.23656557377049187, "calib/final_conf_rate": 0.953125, "calib/format_rate": 0.9453125, "calib/frac_conf_gt_0.9": 0.05327868852459016, "calib/gap": -0.05966184397163132, "calib/mean_conf": 0.5716311475409837, "calib/mu_c": 0.5486466666666666, "calib/mu_w": 0.6083085106382979, "calib/nonempty_final_conf_rate": 0.953125, "calib/nonempty_reasoning_rate": 0.98828125, "calib/nonempty_step_conf_rate": 0.96875, "calib/pce": 0.09672131147540984, "calib/std_conf": 0.21530094076757877, "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": 2554.0, "completions/max_terminated_length": 2554.0, "completions/mean_length": 544.015625, "completions/mean_terminated_length": 544.015625, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.08213333333333334, "grad_norm": 0.004641165491193533, "learning_rate": 3.416666666666667e-06, "loss": 0.0562, "num_tokens": 21443706.0, "reward": 1.7132434844970703, "reward_std": 0.5053609609603882, "rewards/accuracy_reward_step": 0.58984375, "rewards/final_brier_reward_step": 0.6508340835571289, "rewards/format_reward_step": 0.9453125, "rewards/stepwise_brier_reward": 0.772452712059021, "step": 77 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5560347581624177, "calib/avg_num_step_conf": 7.27734375, "calib/ece": 0.13735317460317467, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.07936507936507936, "calib/gap": 0.03492792792792787, "calib/mean_conf": 0.608281746031746, "calib/mu_c": 0.6236666666666667, "calib/mu_w": 0.5887387387387388, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.98828125, "calib/pce": 0.09305555555555559, "calib/std_conf": 0.19237555239389473, "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": 1442.0, "completions/max_terminated_length": 1442.0, "completions/mean_length": 575.99609375, "completions/mean_terminated_length": 578.2549438476562, "completions/min_length": 0.0, "completions/min_terminated_length": 219.0, "epoch": 0.0832, "grad_norm": 0.004652938339859247, "learning_rate": 3.3888888888888893e-06, "loss": 0.0204, "num_tokens": 21752689.0, "reward": 1.699843406677246, "reward_std": 0.49800097942352295, "rewards/accuracy_reward_step": 0.55078125, "rewards/final_brier_reward_step": 0.7162840366363525, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.8252766728401184, "step": 78 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.5047686688311689, "calib/avg_num_step_conf": 6.75, "calib/ece": 0.175568, "calib/final_conf_rate": 0.9765625, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.072, "calib/gap": 0.003684253246753255, "calib/mean_conf": 0.570832, "calib/mu_c": 0.5722467532467532, "calib/mu_w": 0.5685625, "calib/nonempty_final_conf_rate": 0.9765625, "calib/nonempty_reasoning_rate": 0.9921875, "calib/nonempty_step_conf_rate": 0.9921875, "calib/pce": 0.06520000000000002, "calib/std_conf": 0.21249264405150595, "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": 2968.0, "completions/max_terminated_length": 2968.0, "completions/mean_length": 581.83984375, "completions/mean_terminated_length": 581.83984375, "completions/min_length": 129.0, "completions/min_terminated_length": 129.0, "epoch": 0.08426666666666667, "grad_norm": 0.0042600310407578945, "learning_rate": 3.3611111111111117e-06, "loss": 0.0447, "num_tokens": 22061520.0, "reward": 1.7738888263702393, "reward_std": 0.41786515712738037, "rewards/accuracy_reward_step": 0.6015625, "rewards/final_brier_reward_step": 0.7011774778366089, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.8318777084350586, "step": 79 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5680065359477124, "calib/avg_num_step_conf": 7.4921875, "calib/ece": 0.1215257936507936, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.06349206349206349, "calib/gap": 0.05572647058823521, "calib/mean_conf": 0.6064107142857144, "calib/mu_c": 0.6289666666666667, "calib/mu_w": 0.5732401960784315, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.06634920634920632, "calib/std_conf": 0.2234966730432347, "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": 1993.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 528.48828125, "completions/mean_terminated_length": 528.48828125, "completions/min_length": 175.0, "completions/min_terminated_length": 175.0, "epoch": 0.08533333333333333, "grad_norm": 0.004351717419922352, "learning_rate": 3.3333333333333333e-06, "loss": 0.0367, "num_tokens": 22352477.0, "reward": 1.7585244178771973, "reward_std": 0.4870692491531372, "rewards/accuracy_reward_step": 0.5859375, "rewards/final_brier_reward_step": 0.7243493795394897, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8253731727600098, "step": 80 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.4275124639202309, "calib/avg_num_step_conf": 6.97265625, "calib/ece": 0.2267330677290837, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.97265625, "calib/frac_conf_gt_0.9": 0.043824701195219126, "calib/gap": -0.03275977433744415, "calib/mean_conf": 0.5855378486055778, "calib/mu_c": 0.5720945945945947, "calib/mu_w": 0.6048543689320388, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.9921875, "calib/pce": 0.1113147410358566, "calib/std_conf": 0.19887865317073367, "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": 2338.0, "completions/max_terminated_length": 2338.0, "completions/mean_length": 531.98046875, "completions/mean_terminated_length": 534.0667114257812, "completions/min_length": 0.0, "completions/min_terminated_length": 201.0, "epoch": 0.0864, "grad_norm": 0.004251690581440926, "learning_rate": 3.3055555555555558e-06, "loss": 0.0213, "num_tokens": 22648416.0, "reward": 1.7306180000305176, "reward_std": 0.36946582794189453, "rewards/accuracy_reward_step": 0.58203125, "rewards/final_brier_reward_step": 0.6845537424087524, "rewards/format_reward_step": 0.97265625, "rewards/stepwise_brier_reward": 0.8004180192947388, "step": 81 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5432692307692308, "calib/avg_num_step_conf": 6.1015625, "calib/ece": 0.12872549019607848, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.054901960784313725, "calib/gap": 0.02402699949057563, "calib/mean_conf": 0.5968431372549019, "calib/mu_c": 0.6066423841059603, "calib/mu_w": 0.5826153846153846, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.98046875, "calib/pce": 0.06670588235294124, "calib/std_conf": 0.19617586873137727, "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": 1397.0, "completions/max_terminated_length": 1397.0, "completions/mean_length": 479.859375, "completions/mean_terminated_length": 481.7412109375, "completions/min_length": 0.0, "completions/min_terminated_length": 155.0, "epoch": 0.08746666666666666, "grad_norm": 0.004824996925890446, "learning_rate": 3.277777777777778e-06, "loss": 0.0205, "num_tokens": 22930316.0, "reward": 1.7568727731704712, "reward_std": 0.5232962369918823, "rewards/accuracy_reward_step": 0.58984375, "rewards/final_brier_reward_step": 0.7164386510848999, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.8188649415969849, "step": 82 }, { "calib/answer_extract_rate": 0.96875, "calib/auroc": 0.4856194690265486, "calib/avg_num_step_conf": 7.125, "calib/ece": 0.18457831325301205, "calib/final_conf_rate": 0.97265625, "calib/format_rate": 0.95703125, "calib/frac_conf_gt_0.9": 0.07228915662650602, "calib/gap": -0.002629685059864495, "calib/mean_conf": 0.6091566265060241, "calib/mu_c": 0.6079632352941177, "calib/mu_w": 0.6105929203539822, "calib/nonempty_final_conf_rate": 0.97265625, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.984375, "calib/pce": 0.12377510040160644, "calib/std_conf": 0.2203123457389646, "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": 2361.0, "completions/max_terminated_length": 2361.0, "completions/mean_length": 596.3359375, "completions/mean_terminated_length": 596.3359375, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.08853333333333334, "grad_norm": 0.004141851328313351, "learning_rate": 3.2500000000000002e-06, "loss": 0.0516, "num_tokens": 23243746.0, "reward": 1.640852689743042, "reward_std": 0.4232838451862335, "rewards/accuracy_reward_step": 0.53125, "rewards/final_brier_reward_step": 0.6676397323608398, "rewards/format_reward_step": 0.95703125, "rewards/stepwise_brier_reward": 0.7942085266113281, "step": 83 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5440404557675074, "calib/avg_num_step_conf": 6.75390625, "calib/ece": 0.14584980237154152, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.05928853754940711, "calib/gap": 0.030170336704647305, "calib/mean_conf": 0.6058498023715415, "calib/mu_c": 0.618609589041096, "calib/mu_w": 0.5884392523364487, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.08731225296442689, "calib/std_conf": 0.20479879543257298, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2668.0, "completions/max_terminated_length": 2668.0, "completions/mean_length": 512.37109375, "completions/mean_terminated_length": 512.37109375, "completions/min_length": 147.0, "completions/min_terminated_length": 147.0, "epoch": 0.0896, "grad_norm": 0.0048288702964782715, "learning_rate": 3.2222222222222227e-06, "loss": 0.0315, "num_tokens": 23534337.0, "reward": 1.7381473779678345, "reward_std": 0.4058360457420349, "rewards/accuracy_reward_step": 0.5703125, "rewards/final_brier_reward_step": 0.7193666696548462, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8347854614257812, "step": 84 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.3971710108073745, "calib/avg_num_step_conf": 6.26953125, "calib/ece": 0.2807055335968379, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.07509881422924901, "calib/gap": -0.07377692307692296, "calib/mean_conf": 0.6100454545454546, "calib/mu_c": 0.5683454545454546, "calib/mu_w": 0.6421223776223776, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.22798418972332013, "calib/std_conf": 0.20210970025665978, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2066.0, "completions/max_terminated_length": 2066.0, "completions/mean_length": 533.12890625, "completions/mean_terminated_length": 533.12890625, "completions/min_length": 137.0, "completions/min_terminated_length": 137.0, "epoch": 0.09066666666666667, "grad_norm": 0.0045362371020019054, "learning_rate": 3.1944444444444443e-06, "loss": 0.002, "num_tokens": 23832146.0, "reward": 1.5035837888717651, "reward_std": 0.4948905408382416, "rewards/accuracy_reward_step": 0.4296875, "rewards/final_brier_reward_step": 0.6388517022132874, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8207959532737732, "step": 85 }, { "calib/answer_extract_rate": 0.96875, "calib/auroc": 0.5581788617886179, "calib/avg_num_step_conf": 6.2265625, "calib/ece": 0.16031250000000002, "calib/final_conf_rate": 0.96875, "calib/format_rate": 0.96484375, "calib/frac_conf_gt_0.9": 0.024193548387096774, "calib/gap": 0.02073551219512193, "calib/mean_conf": 0.5972681451612903, "calib/mu_c": 0.6077195121951219, "calib/mu_w": 0.586984, "calib/nonempty_final_conf_rate": 0.96875, "calib/nonempty_reasoning_rate": 0.98046875, "calib/nonempty_step_conf_rate": 0.9765625, "calib/pce": 0.13080645161290325, "calib/std_conf": 0.1918157867374864, "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": 2063.0, "completions/max_terminated_length": 2063.0, "completions/mean_length": 531.984375, "completions/mean_terminated_length": 536.1732177734375, "completions/min_length": 0.0, "completions/min_terminated_length": 131.0, "epoch": 0.09173333333333333, "grad_norm": 0.004984323401004076, "learning_rate": 3.1666666666666667e-06, "loss": 0.0044, "num_tokens": 24127350.0, "reward": 1.5809977054595947, "reward_std": 0.454933226108551, "rewards/accuracy_reward_step": 0.48046875, "rewards/final_brier_reward_step": 0.6893277764320374, "rewards/format_reward_step": 0.96484375, "rewards/stepwise_brier_reward": 0.8221626281738281, "step": 86 }, { "calib/answer_extract_rate": 0.97265625, "calib/auroc": 0.41233098177542626, "calib/avg_num_step_conf": 5.79296875, "calib/ece": 0.20172610441767072, "calib/final_conf_rate": 0.97265625, "calib/format_rate": 0.96875, "calib/frac_conf_gt_0.9": 0.04417670682730924, "calib/gap": -0.06079312169312168, "calib/mean_conf": 0.6040570281124497, "calib/mu_c": 0.5842809523809523, "calib/mu_w": 0.645074074074074, "calib/nonempty_final_conf_rate": 0.97265625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.06554216867469884, "calib/std_conf": 0.18658293299318823, "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": 2762.0, "completions/max_terminated_length": 2762.0, "completions/mean_length": 454.35546875, "completions/mean_terminated_length": 457.9330749511719, "completions/min_length": 0.0, "completions/min_terminated_length": 138.0, "epoch": 0.0928, "grad_norm": 0.005657955538481474, "learning_rate": 3.138888888888889e-06, "loss": 0.062, "num_tokens": 24402665.0, "reward": 1.8470996618270874, "reward_std": 0.497073233127594, "rewards/accuracy_reward_step": 0.65625, "rewards/final_brier_reward_step": 0.6915764808654785, "rewards/format_reward_step": 0.96875, "rewards/stepwise_brier_reward": 0.8218224048614502, "step": 87 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5025510204081634, "calib/avg_num_step_conf": 6.4453125, "calib/ece": 0.17941764705882352, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.10980392156862745, "calib/gap": 0.009866496598639474, "calib/mean_conf": 0.6008960784313725, "calib/mu_c": 0.6050748299319728, "calib/mu_w": 0.5952083333333333, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.10192156862745094, "calib/std_conf": 0.22941628825773167, "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": 1557.0, "completions/max_terminated_length": 1557.0, "completions/mean_length": 512.13671875, "completions/mean_terminated_length": 514.1451416015625, "completions/min_length": 0.0, "completions/min_terminated_length": 144.0, "epoch": 0.09386666666666667, "grad_norm": 0.00500339362770319, "learning_rate": 3.1111111111111116e-06, "loss": 0.04, "num_tokens": 24697124.0, "reward": 1.7386332750320435, "reward_std": 0.4612874984741211, "rewards/accuracy_reward_step": 0.57421875, "rewards/final_brier_reward_step": 0.7017440795898438, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.8231015801429749, "step": 88 }, { "calib/answer_extract_rate": 0.96484375, "calib/auroc": 0.5398230088495575, "calib/avg_num_step_conf": 6.5078125, "calib/ece": 0.24250607287449394, "calib/final_conf_rate": 0.96484375, "calib/format_rate": 0.953125, "calib/frac_conf_gt_0.9": 0.05263157894736842, "calib/gap": 0.023540021133271805, "calib/mean_conf": 0.6186275303643725, "calib/mu_c": 0.6313982300884956, "calib/mu_w": 0.6078582089552238, "calib/nonempty_final_conf_rate": 0.96484375, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.984375, "calib/pce": 0.20182186234817817, "calib/std_conf": 0.21626131041815927, "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": 3068.0, "completions/max_terminated_length": 3068.0, "completions/mean_length": 559.08203125, "completions/mean_terminated_length": 559.08203125, "completions/min_length": 147.0, "completions/min_terminated_length": 147.0, "epoch": 0.09493333333333333, "grad_norm": 0.005031867418438196, "learning_rate": 3.0833333333333336e-06, "loss": 0.084, "num_tokens": 25002641.0, "reward": 1.4930341243743896, "reward_std": 0.41416651010513306, "rewards/accuracy_reward_step": 0.44140625, "rewards/final_brier_reward_step": 0.6576842069625854, "rewards/format_reward_step": 0.953125, "rewards/stepwise_brier_reward": 0.759764552116394, "step": 89 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5396780303030304, "calib/avg_num_step_conf": 7.00390625, "calib/ece": 0.16229330708661419, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.08267716535433071, "calib/gap": 0.03228674242424234, "calib/mean_conf": 0.6441633858267717, "calib/mu_c": 0.6581458333333334, "calib/mu_w": 0.6258590909090911, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.11976377952755907, "calib/std_conf": 0.20476278256550734, "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": 2047.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 482.56640625, "completions/mean_terminated_length": 482.56640625, "completions/min_length": 123.0, "completions/min_terminated_length": 123.0, "epoch": 0.096, "grad_norm": 0.005818725563585758, "learning_rate": 3.055555555555556e-06, "loss": -0.0018, "num_tokens": 25283002.0, "reward": 1.7255547046661377, "reward_std": 0.5528163909912109, "rewards/accuracy_reward_step": 0.5625, "rewards/final_brier_reward_step": 0.7167966365814209, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.826046884059906, "step": 90 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.5589280831282473, "calib/avg_num_step_conf": 6.78125, "calib/ece": 0.11946414342629487, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.08366533864541832, "calib/gap": 0.038767637407711164, "calib/mean_conf": 0.6494601593625497, "calib/mu_c": 0.6636698113207548, "calib/mu_w": 0.6249021739130436, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.9921875, "calib/pce": 0.06772908366533867, "calib/std_conf": 0.18501397709226325, "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": 2457.0, "completions/max_terminated_length": 2457.0, "completions/mean_length": 536.078125, "completions/mean_terminated_length": 536.078125, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.09706666666666666, "grad_norm": 0.0048302579671144485, "learning_rate": 3.0277777777777776e-06, "loss": 0.0352, "num_tokens": 25581454.0, "reward": 1.8105878829956055, "reward_std": 0.4493829607963562, "rewards/accuracy_reward_step": 0.62109375, "rewards/final_brier_reward_step": 0.7332277297973633, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.829436182975769, "step": 91 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5373383790602334, "calib/avg_num_step_conf": 6.45703125, "calib/ece": 0.13286718749999998, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.046875, "calib/gap": 0.020359129612109728, "calib/mean_conf": 0.6113515624999999, "calib/mu_c": 0.6197019867549668, "calib/mu_w": 0.5993428571428571, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.07718749999999999, "calib/std_conf": 0.1758677188366546, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1458.0, "completions/max_terminated_length": 1458.0, "completions/mean_length": 449.94921875, "completions/mean_terminated_length": 451.7137451171875, "completions/min_length": 0.0, "completions/min_terminated_length": 177.0, "epoch": 0.09813333333333334, "grad_norm": 0.005348356440663338, "learning_rate": 3e-06, "loss": -0.0385, "num_tokens": 25856865.0, "reward": 1.7782703638076782, "reward_std": 0.3455822467803955, "rewards/accuracy_reward_step": 0.58984375, "rewards/final_brier_reward_step": 0.7365307807922363, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.8374884128570557, "step": 92 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.45004985666209646, "calib/avg_num_step_conf": 7.171875, "calib/ece": 0.2897031372549019, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.09411764705882353, "calib/gap": -0.017236576093730527, "calib/mean_conf": 0.6701400000000001, "calib/mu_c": 0.660541592920354, "calib/mu_w": 0.6777781690140845, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.25835294117647056, "calib/std_conf": 0.18891003452727978, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2557.0, "completions/max_terminated_length": 2557.0, "completions/mean_length": 533.30859375, "completions/mean_terminated_length": 533.30859375, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.0992, "grad_norm": 0.00487908348441124, "learning_rate": 2.9722222222222225e-06, "loss": 0.0259, "num_tokens": 26152672.0, "reward": 1.526746153831482, "reward_std": 0.5700970888137817, "rewards/accuracy_reward_step": 0.44140625, "rewards/final_brier_reward_step": 0.6549409031867981, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8114188313484192, "step": 93 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5477315689981097, "calib/avg_num_step_conf": 5.8828125, "calib/ece": 0.1766798418972332, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.03162055335968379, "calib/gap": 0.019027536231884135, "calib/mean_conf": 0.6361264822134387, "calib/mu_c": 0.6447753623188407, "calib/mu_w": 0.6257478260869566, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.13367588932806326, "calib/std_conf": 0.1764306843318111, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 1694.0, "completions/max_terminated_length": 1694.0, "completions/mean_length": 417.8046875, "completions/mean_terminated_length": 421.094482421875, "completions/min_length": 0.0, "completions/min_terminated_length": 111.0, "epoch": 0.10026666666666667, "grad_norm": 0.006482561584562063, "learning_rate": 2.944444444444445e-06, "loss": -0.0073, "num_tokens": 26421814.0, "reward": 1.6926339864730835, "reward_std": 0.3536980152130127, "rewards/accuracy_reward_step": 0.54296875, "rewards/final_brier_reward_step": 0.7136893272399902, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8224716186523438, "step": 94 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5479842101857244, "calib/avg_num_step_conf": 6.57421875, "calib/ece": 0.11718503937007875, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.07086614173228346, "calib/gap": 0.013154403675661785, "calib/mean_conf": 0.649507874015748, "calib/mu_c": 0.6547385620915034, "calib/mu_w": 0.6415841584158416, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.08216535433070869, "calib/std_conf": 0.16144584772927267, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 1446.0, "completions/max_terminated_length": 1446.0, "completions/mean_length": 496.1328125, "completions/mean_terminated_length": 500.03936767578125, "completions/min_length": 0.0, "completions/min_terminated_length": 101.0, "epoch": 0.10133333333333333, "grad_norm": 0.005486312322318554, "learning_rate": 2.916666666666667e-06, "loss": 0.0288, "num_tokens": 26708456.0, "reward": 1.7844123840332031, "reward_std": 0.506519079208374, "rewards/accuracy_reward_step": 0.59765625, "rewards/final_brier_reward_step": 0.7327226400375366, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.834614634513855, "step": 95 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.47741886099203923, "calib/avg_num_step_conf": 6.66796875, "calib/ece": 0.18236274509803924, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.06666666666666667, "calib/gap": -0.02285727954684602, "calib/mean_conf": 0.6447745098039216, "calib/mu_c": 0.6384103260869566, "calib/mu_w": 0.6612676056338026, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.052784313725490195, "calib/std_conf": 0.19463172145698326, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 975.0, "completions/max_terminated_length": 975.0, "completions/mean_length": 440.90625, "completions/mean_terminated_length": 442.63531494140625, "completions/min_length": 0.0, "completions/min_terminated_length": 126.0, "epoch": 0.1024, "grad_norm": 0.0060312505811452866, "learning_rate": 2.888888888888889e-06, "loss": 0.0157, "num_tokens": 26980648.0, "reward": 1.9723178148269653, "reward_std": 0.3152076005935669, "rewards/accuracy_reward_step": 0.71875, "rewards/final_brier_reward_step": 0.7432148456573486, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.841369092464447, "step": 96 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.4821138211382114, "calib/avg_num_step_conf": 6.50390625, "calib/ece": 0.25586956521739135, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.06324110671936758, "calib/gap": -0.009362914321450866, "calib/mean_conf": 0.6590711462450594, "calib/mu_c": 0.654260162601626, "calib/mu_w": 0.6636230769230769, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.21438735177865623, "calib/std_conf": 0.17423228236708665, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2313.0, "completions/max_terminated_length": 2313.0, "completions/mean_length": 473.13671875, "completions/mean_terminated_length": 476.8622131347656, "completions/min_length": 0.0, "completions/min_terminated_length": 165.0, "epoch": 0.10346666666666667, "grad_norm": 0.006380012258887291, "learning_rate": 2.861111111111111e-06, "loss": 0.0316, "num_tokens": 27260347.0, "reward": 1.5849992036819458, "reward_std": 0.4791141152381897, "rewards/accuracy_reward_step": 0.48046875, "rewards/final_brier_reward_step": 0.6772300004959106, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8033915758132935, "step": 97 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5214604591836735, "calib/avg_num_step_conf": 6.85546875, "calib/ece": 0.20126190476190473, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.12698412698412698, "calib/gap": 0.007683928571428633, "calib/mean_conf": 0.6912777777777778, "calib/mu_c": 0.6946928571428572, "calib/mu_w": 0.6870089285714286, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.16849206349206344, "calib/std_conf": 0.1930189413720712, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2315.0, "completions/max_terminated_length": 2315.0, "completions/mean_length": 535.96875, "completions/mean_terminated_length": 535.96875, "completions/min_length": 96.0, "completions/min_terminated_length": 96.0, "epoch": 0.10453333333333334, "grad_norm": 0.00536646181717515, "learning_rate": 2.8333333333333335e-06, "loss": 0.0359, "num_tokens": 27557243.0, "reward": 1.6928212642669678, "reward_std": 0.4175313413143158, "rewards/accuracy_reward_step": 0.546875, "rewards/final_brier_reward_step": 0.6902477741241455, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8310375809669495, "step": 98 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5127483336568951, "calib/avg_num_step_conf": 7.0859375, "calib/ece": 0.3450937007874016, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.0984251968503937, "calib/gap": -0.0041598524558338346, "calib/mean_conf": 0.6790007874015749, "calib/mu_c": 0.6764950495049504, "calib/mu_w": 0.6806549019607843, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.3132283464566929, "calib/std_conf": 0.21706684027963896, "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": 1708.0, "completions/max_terminated_length": 1708.0, "completions/mean_length": 568.1015625, "completions/mean_terminated_length": 570.3294677734375, "completions/min_length": 0.0, "completions/min_terminated_length": 127.0, "epoch": 0.1056, "grad_norm": 0.005416756495833397, "learning_rate": 2.805555555555556e-06, "loss": 0.0491, "num_tokens": 27861981.0, "reward": 1.437119722366333, "reward_std": 0.5040462017059326, "rewards/accuracy_reward_step": 0.39453125, "rewards/final_brier_reward_step": 0.6272630095481873, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.7696532011032104, "step": 99 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5930734576398449, "calib/avg_num_step_conf": 6.75390625, "calib/ece": 0.2713320158102766, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.15019762845849802, "calib/gap": 0.038916781379051524, "calib/mean_conf": 0.6681936758893281, "calib/mu_c": 0.6883442622950819, "calib/mu_w": 0.6494274809160304, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.2286561264822134, "calib/std_conf": 0.22656343953497585, "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": 2497.0, "completions/max_terminated_length": 2497.0, "completions/mean_length": 535.96875, "completions/mean_terminated_length": 538.0706176757812, "completions/min_length": 0.0, "completions/min_terminated_length": 149.0, "epoch": 0.10666666666666667, "grad_norm": 0.0061225686222314835, "learning_rate": 2.7777777777777783e-06, "loss": 0.0542, "num_tokens": 28160101.0, "reward": 1.5802373886108398, "reward_std": 0.4150017499923706, "rewards/accuracy_reward_step": 0.4765625, "rewards/final_brier_reward_step": 0.6758168935775757, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8091955184936523, "step": 100 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5235927152317881, "calib/avg_num_step_conf": 7.4921875, "calib/ece": 0.35295882352941177, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.1843137254901961, "calib/gap": 0.009206444218033516, "calib/mean_conf": 0.6984137254901961, "calib/mu_c": 0.7038653846153846, "calib/mu_w": 0.6946589403973511, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.32176470588235295, "calib/std_conf": 0.2336589739045362, "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": 1262.0, "completions/max_terminated_length": 1262.0, "completions/mean_length": 526.68359375, "completions/mean_terminated_length": 528.7490234375, "completions/min_length": 0.0, "completions/min_terminated_length": 124.0, "epoch": 0.10773333333333333, "grad_norm": 0.006173782516270876, "learning_rate": 2.7500000000000004e-06, "loss": -0.0296, "num_tokens": 28455428.0, "reward": 1.4512096643447876, "reward_std": 0.4737076759338379, "rewards/accuracy_reward_step": 0.40625, "rewards/final_brier_reward_step": 0.6192189455032349, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.7637447118759155, "step": 101 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5934290913817757, "calib/avg_num_step_conf": 6.66796875, "calib/ece": 0.09631176470588242, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.08235294117647059, "calib/gap": 0.05825776680098793, "calib/mean_conf": 0.669021568627451, "calib/mu_c": 0.6914108280254777, "calib/mu_w": 0.6331530612244898, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.07482352941176476, "calib/std_conf": 0.18855365004835895, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1783.0, "completions/max_terminated_length": 1783.0, "completions/mean_length": 449.5546875, "completions/mean_terminated_length": 451.31768798828125, "completions/min_length": 0.0, "completions/min_terminated_length": 94.0, "epoch": 0.1088, "grad_norm": 0.007125023752450943, "learning_rate": 2.7222222222222224e-06, "loss": -0.0151, "num_tokens": 28730714.0, "reward": 1.815610408782959, "reward_std": 0.3154347836971283, "rewards/accuracy_reward_step": 0.61328125, "rewards/final_brier_reward_step": 0.7496160268783569, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8409508466720581, "step": 102 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.543672537508154, "calib/avg_num_step_conf": 6.8828125, "calib/ece": 0.1759023904382471, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.13147410358565736, "calib/gap": 0.024448238747553974, "calib/mean_conf": 0.6975637450199204, "calib/mu_c": 0.707791095890411, "calib/mu_w": 0.683342857142857, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.14589641434262957, "calib/std_conf": 0.18706129532185103, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2515.0, "completions/max_terminated_length": 2515.0, "completions/mean_length": 583.5703125, "completions/mean_terminated_length": 585.85888671875, "completions/min_length": 0.0, "completions/min_terminated_length": 141.0, "epoch": 0.10986666666666667, "grad_norm": 0.006434804294258356, "learning_rate": 2.6944444444444444e-06, "loss": -0.0001, "num_tokens": 29038164.0, "reward": 1.7238279581069946, "reward_std": 0.3587304949760437, "rewards/accuracy_reward_step": 0.5703125, "rewards/final_brier_reward_step": 0.7060806155204773, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.8064184188842773, "step": 103 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5021584084084084, "calib/avg_num_step_conf": 7.1171875, "calib/ece": 0.30296862745098035, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.09411764705882353, "calib/gap": -0.011483858858858942, "calib/mean_conf": 0.7084039215686275, "calib/mu_c": 0.7019189189189188, "calib/mu_w": 0.7134027777777777, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.2880392156862745, "calib/std_conf": 0.15928663762474796, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2477.0, "completions/max_terminated_length": 2477.0, "completions/mean_length": 508.421875, "completions/mean_terminated_length": 508.421875, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.11093333333333333, "grad_norm": 0.007167415227741003, "learning_rate": 2.666666666666667e-06, "loss": 0.0198, "num_tokens": 29328504.0, "reward": 1.5080779790878296, "reward_std": 0.392715722322464, "rewards/accuracy_reward_step": 0.43359375, "rewards/final_brier_reward_step": 0.6460463404655457, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.7925153970718384, "step": 104 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.4136380550731828, "calib/avg_num_step_conf": 7.33984375, "calib/ece": 0.24572549019607845, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.08235294117647059, "calib/gap": -0.04938972959563381, "calib/mean_conf": 0.6672156862745099, "calib/mu_c": 0.6447482014388489, "calib/mu_w": 0.6941379310344827, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.183921568627451, "calib/std_conf": 0.17820069801644614, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1962.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 513.0, "completions/mean_terminated_length": 515.0117797851562, "completions/min_length": 0.0, "completions/min_terminated_length": 170.0, "epoch": 0.112, "grad_norm": 0.0071303932927548885, "learning_rate": 2.6388888888888893e-06, "loss": 0.0273, "num_tokens": 29619096.0, "reward": 1.6805466413497925, "reward_std": 0.4850854277610779, "rewards/accuracy_reward_step": 0.54296875, "rewards/final_brier_reward_step": 0.6782119870185852, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.793974757194519, "step": 105 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.466280950603459, "calib/avg_num_step_conf": 6.625, "calib/ece": 0.13909098039215687, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.054901960784313725, "calib/gap": -0.004165098917506382, "calib/mean_conf": 0.6504776470588236, "calib/mu_c": 0.6486156028368795, "calib/mu_w": 0.6527807017543859, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.11831372549019609, "calib/std_conf": 0.17081772721740143, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2442.0, "completions/max_terminated_length": 2442.0, "completions/mean_length": 482.05859375, "completions/mean_terminated_length": 482.05859375, "completions/min_length": 154.0, "completions/min_terminated_length": 154.0, "epoch": 0.11306666666666666, "grad_norm": 0.0068705882877111435, "learning_rate": 2.6111111111111113e-06, "loss": 0.0173, "num_tokens": 29900591.0, "reward": 1.7094638347625732, "reward_std": 0.30682140588760376, "rewards/accuracy_reward_step": 0.55078125, "rewards/final_brier_reward_step": 0.7092701196670532, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8317103981971741, "step": 106 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5721975727521857, "calib/avg_num_step_conf": 7.41796875, "calib/ece": 0.07635686274509804, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.050980392156862744, "calib/gap": 0.03128357040323637, "calib/mean_conf": 0.6742392156862744, "calib/mu_c": 0.6861392405063291, "calib/mu_w": 0.6548556701030928, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.06549411764705884, "calib/std_conf": 0.1267651589394873, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1175.0, "completions/max_terminated_length": 1175.0, "completions/mean_length": 477.890625, "completions/mean_terminated_length": 479.7647399902344, "completions/min_length": 0.0, "completions/min_terminated_length": 137.0, "epoch": 0.11413333333333334, "grad_norm": 0.00758032500743866, "learning_rate": 2.5833333333333337e-06, "loss": -0.0045, "num_tokens": 30181051.0, "reward": 1.8239885568618774, "reward_std": 0.3716495931148529, "rewards/accuracy_reward_step": 0.6171875, "rewards/final_brier_reward_step": 0.7570300102233887, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8436118364334106, "step": 107 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5208333333333334, "calib/avg_num_step_conf": 7.375, "calib/ece": 0.17114541832669328, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.043824701195219126, "calib/gap": 0.006283901515151458, "calib/mean_conf": 0.6542729083665338, "calib/mu_c": 0.6561505681818182, "calib/mu_w": 0.6498666666666667, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.062111553784860565, "calib/std_conf": 0.15603809367075067, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2487.0, "completions/max_terminated_length": 2487.0, "completions/mean_length": 540.9453125, "completions/mean_terminated_length": 543.0667114257812, "completions/min_length": 0.0, "completions/min_terminated_length": 116.0, "epoch": 0.1152, "grad_norm": 0.006146944127976894, "learning_rate": 2.5555555555555557e-06, "loss": 0.0056, "num_tokens": 30476269.0, "reward": 1.9283292293548584, "reward_std": 0.4546602964401245, "rewards/accuracy_reward_step": 0.6953125, "rewards/final_brier_reward_step": 0.7515912055969238, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.8289128541946411, "step": 108 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5315239168544953, "calib/avg_num_step_conf": 7.484375, "calib/ece": 0.16493873517786567, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.039525691699604744, "calib/gap": -0.007344696969696973, "calib/mean_conf": 0.6452588932806324, "calib/mu_c": 0.6417462121212121, "calib/mu_w": 0.649090909090909, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.14422924901185774, "calib/std_conf": 0.15751775472409651, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2387.0, "completions/max_terminated_length": 2387.0, "completions/mean_length": 520.86328125, "completions/mean_terminated_length": 522.9058837890625, "completions/min_length": 0.0, "completions/min_terminated_length": 160.0, "epoch": 0.11626666666666667, "grad_norm": 0.006327529903501272, "learning_rate": 2.5277777777777778e-06, "loss": 0.0387, "num_tokens": 30767714.0, "reward": 1.6458185911178589, "reward_std": 0.2906777858734131, "rewards/accuracy_reward_step": 0.515625, "rewards/final_brier_reward_step": 0.6984561681747437, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8145054578781128, "step": 109 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.4587017813661412, "calib/avg_num_step_conf": 6.0859375, "calib/ece": 0.1278039215686274, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.00784313725490196, "calib/gap": -0.01728758169934619, "calib/mean_conf": 0.6134509803921568, "calib/mu_c": 0.6065359477124184, "calib/mu_w": 0.6238235294117646, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.07062745098039212, "calib/std_conf": 0.12247709889583763, "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": 1586.0, "completions/max_terminated_length": 1586.0, "completions/mean_length": 463.87890625, "completions/mean_terminated_length": 465.69805908203125, "completions/min_length": 0.0, "completions/min_terminated_length": 96.0, "epoch": 0.11733333333333333, "grad_norm": 0.007411710452288389, "learning_rate": 2.5e-06, "loss": 0.0186, "num_tokens": 31044891.0, "reward": 1.7877941131591797, "reward_std": 0.4205990433692932, "rewards/accuracy_reward_step": 0.59765625, "rewards/final_brier_reward_step": 0.7311433553695679, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.8497209548950195, "step": 110 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.4844827586206896, "calib/avg_num_step_conf": 6.84765625, "calib/ece": 0.1450996015936256, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.0199203187250996, "calib/gap": -0.0036762452107280685, "calib/mean_conf": 0.6241434262948207, "calib/mu_c": 0.6224444444444445, "calib/mu_w": 0.6261206896551725, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.11569721115537854, "calib/std_conf": 0.14558702216380226, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2715.0, "completions/max_terminated_length": 2715.0, "completions/mean_length": 533.65234375, "completions/mean_terminated_length": 537.8543090820312, "completions/min_length": 0.0, "completions/min_terminated_length": 105.0, "epoch": 0.1184, "grad_norm": 0.006978228688240051, "learning_rate": 2.4722222222222226e-06, "loss": 0.0345, "num_tokens": 31342418.0, "reward": 1.6597996950149536, "reward_std": 0.4123978018760681, "rewards/accuracy_reward_step": 0.52734375, "rewards/final_brier_reward_step": 0.7048890590667725, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.8171224594116211, "step": 111 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.5674880000000001, "calib/avg_num_step_conf": 6.48828125, "calib/ece": 0.14463999999999996, "calib/final_conf_rate": 0.9765625, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.02, "calib/gap": 0.025439999999999796, "calib/mean_conf": 0.6338400000000001, "calib/mu_c": 0.6465599999999999, "calib/mu_w": 0.6211200000000001, "calib/nonempty_final_conf_rate": 0.9765625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.13923999999999995, "calib/std_conf": 0.12414690652609915, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2419.0, "completions/max_terminated_length": 2419.0, "completions/mean_length": 557.21875, "completions/mean_terminated_length": 557.21875, "completions/min_length": 165.0, "completions/min_terminated_length": 165.0, "epoch": 0.11946666666666667, "grad_norm": 0.005996756721287966, "learning_rate": 2.4444444444444447e-06, "loss": 0.0632, "num_tokens": 31646490.0, "reward": 1.6017860174179077, "reward_std": 0.3949372172355652, "rewards/accuracy_reward_step": 0.48828125, "rewards/final_brier_reward_step": 0.7122992277145386, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.8120320439338684, "step": 112 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5166979362101314, "calib/avg_num_step_conf": 6.83203125, "calib/ece": 0.19297233201581035, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.05533596837944664, "calib/gap": 0.006614133833645841, "calib/mean_conf": 0.648292490118577, "calib/mu_c": 0.6516910569105688, "calib/mu_w": 0.645076923076923, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.17754940711462458, "calib/std_conf": 0.15634266555881815, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1341.0, "completions/max_terminated_length": 1341.0, "completions/mean_length": 474.24609375, "completions/mean_terminated_length": 476.10589599609375, "completions/min_length": 0.0, "completions/min_terminated_length": 128.0, "epoch": 0.12053333333333334, "grad_norm": 0.007260980550199747, "learning_rate": 2.4166666666666667e-06, "loss": 0.0083, "num_tokens": 31926601.0, "reward": 1.5960609912872314, "reward_std": 0.40254390239715576, "rewards/accuracy_reward_step": 0.48046875, "rewards/final_brier_reward_step": 0.6945323944091797, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8303369283676147, "step": 113 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5016359918200409, "calib/avg_num_step_conf": 6.09765625, "calib/ece": 0.08928853754940705, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.011857707509881422, "calib/gap": -0.0018036809815951127, "calib/mean_conf": 0.6268379446640316, "calib/mu_c": 0.6261963190184049, "calib/mu_w": 0.628, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.0359288537549407, "calib/std_conf": 0.11570492584393413, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2535.0, "completions/max_terminated_length": 2535.0, "completions/mean_length": 469.8828125, "completions/mean_terminated_length": 469.8828125, "completions/min_length": 144.0, "completions/min_terminated_length": 144.0, "epoch": 0.1216, "grad_norm": 0.007122975308448076, "learning_rate": 2.388888888888889e-06, "loss": 0.0547, "num_tokens": 32205419.0, "reward": 1.8431209325790405, "reward_std": 0.3087952733039856, "rewards/accuracy_reward_step": 0.63671875, "rewards/final_brier_reward_step": 0.7474324107170105, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8281762599945068, "step": 114 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5428331466965286, "calib/avg_num_step_conf": 5.9375, "calib/ece": 0.08803921568627457, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.03137254901960784, "calib/gap": 0.018712206047032498, "calib/mean_conf": 0.6226274509803921, "calib/mu_c": 0.6309929078014185, "calib/mu_w": 0.612280701754386, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.07886274509803924, "calib/std_conf": 0.1361852775988615, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2637.0, "completions/max_terminated_length": 2637.0, "completions/mean_length": 443.52734375, "completions/mean_terminated_length": 443.52734375, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.12266666666666666, "grad_norm": 0.007569754030555487, "learning_rate": 2.361111111111111e-06, "loss": 0.0546, "num_tokens": 32477730.0, "reward": 1.719649076461792, "reward_std": 0.3827745318412781, "rewards/accuracy_reward_step": 0.55078125, "rewards/final_brier_reward_step": 0.7357659935951233, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8459555506706238, "step": 115 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.45060080106809086, "calib/avg_num_step_conf": 6.65234375, "calib/ece": 0.15657086614173227, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.027559055118110236, "calib/gap": -0.029352342806281406, "calib/mean_conf": 0.6266574803149606, "calib/mu_c": 0.6142925170068027, "calib/mu_w": 0.6436448598130841, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.10224409448818896, "calib/std_conf": 0.1551456415173825, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1478.0, "completions/max_terminated_length": 1478.0, "completions/mean_length": 529.015625, "completions/mean_terminated_length": 531.0902099609375, "completions/min_length": 0.0, "completions/min_terminated_length": 100.0, "epoch": 0.12373333333333333, "grad_norm": 0.006859073415398598, "learning_rate": 2.3333333333333336e-06, "loss": 0.0055, "num_tokens": 32771182.0, "reward": 1.7464134693145752, "reward_std": 0.3535025715827942, "rewards/accuracy_reward_step": 0.57421875, "rewards/final_brier_reward_step": 0.7099315524101257, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.8460346460342407, "step": 116 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.5032380952380953, "calib/avg_num_step_conf": 6.109375, "calib/ece": 0.13629482071713148, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.97265625, "calib/frac_conf_gt_0.9": 0.01593625498007968, "calib/gap": 0.014129523809523703, "calib/mean_conf": 0.6232270916334661, "calib/mu_c": 0.6303200000000001, "calib/mu_w": 0.6161904761904764, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 0.9921875, "calib/nonempty_step_conf_rate": 0.984375, "calib/pce": 0.1307569721115538, "calib/std_conf": 0.1378446214563694, "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": 2534.0, "completions/max_terminated_length": 2534.0, "completions/mean_length": 482.9765625, "completions/mean_terminated_length": 486.7795104980469, "completions/min_length": 0.0, "completions/min_terminated_length": 185.0, "epoch": 0.1248, "grad_norm": 0.006722697056829929, "learning_rate": 2.305555555555556e-06, "loss": 0.0244, "num_tokens": 33054928.0, "reward": 1.597266674041748, "reward_std": 0.4738282263278961, "rewards/accuracy_reward_step": 0.48828125, "rewards/final_brier_reward_step": 0.7031722664833069, "rewards/format_reward_step": 0.97265625, "rewards/stepwise_brier_reward": 0.8108941316604614, "step": 117 }, { "calib/answer_extract_rate": 0.96875, "calib/auroc": 0.5215608465608466, "calib/avg_num_step_conf": 6.5390625, "calib/ece": 0.12467741935483871, "calib/final_conf_rate": 0.96875, "calib/format_rate": 0.96875, "calib/frac_conf_gt_0.9": 0.024193548387096774, "calib/gap": 0.025560846560846562, "calib/mean_conf": 0.600725806451613, "calib/mu_c": 0.6118571428571429, "calib/mu_w": 0.5862962962962963, "calib/nonempty_final_conf_rate": 0.96875, "calib/nonempty_reasoning_rate": 0.98828125, "calib/nonempty_step_conf_rate": 0.98828125, "calib/pce": 0.0804435483870968, "calib/std_conf": 0.1727279857925824, "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": 2598.0, "completions/max_terminated_length": 2598.0, "completions/mean_length": 524.45703125, "completions/mean_terminated_length": 526.5137329101562, "completions/min_length": 0.0, "completions/min_terminated_length": 139.0, "epoch": 0.12586666666666665, "grad_norm": 0.007416529580950737, "learning_rate": 2.277777777777778e-06, "loss": -0.018, "num_tokens": 33346701.0, "reward": 1.681166172027588, "reward_std": 0.34025275707244873, "rewards/accuracy_reward_step": 0.546875, "rewards/final_brier_reward_step": 0.7125968933105469, "rewards/format_reward_step": 0.96875, "rewards/stepwise_brier_reward": 0.793317437171936, "step": 118 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5117266734279919, "calib/avg_num_step_conf": 6.46484375, "calib/ece": 0.11876984126984125, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.007936507936507936, "calib/gap": 0.009959432048681549, "calib/mean_conf": 0.6261507936507936, "calib/mu_c": 0.6307352941176471, "calib/mu_w": 0.6207758620689655, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.10261904761904762, "calib/std_conf": 0.11684622695379758, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2409.0, "completions/max_terminated_length": 2409.0, "completions/mean_length": 524.41015625, "completions/mean_terminated_length": 528.5393676757812, "completions/min_length": 0.0, "completions/min_terminated_length": 93.0, "epoch": 0.12693333333333334, "grad_norm": 0.006316907238215208, "learning_rate": 2.25e-06, "loss": -0.023, "num_tokens": 33639518.0, "reward": 1.6775588989257812, "reward_std": 0.4439692497253418, "rewards/accuracy_reward_step": 0.53125, "rewards/final_brier_reward_step": 0.723902702331543, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.830082893371582, "step": 119 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.49333874349090423, "calib/avg_num_step_conf": 6.1640625, "calib/ece": 0.12428571428571428, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.01984126984126984, "calib/gap": 0.010087238790829844, "calib/mean_conf": 0.6153968253968255, "calib/mu_c": 0.619119496855346, "calib/mu_w": 0.6090322580645161, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 0.98828125, "calib/nonempty_step_conf_rate": 0.98828125, "calib/pce": 0.05436507936507934, "calib/std_conf": 0.13896528511147083, "calib/step_conf_rate": 0.98828125, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1154.0, "completions/max_terminated_length": 1154.0, "completions/mean_length": 449.76953125, "completions/mean_terminated_length": 451.5333557128906, "completions/min_length": 0.0, "completions/min_terminated_length": 179.0, "epoch": 0.128, "grad_norm": 0.00672755716368556, "learning_rate": 2.222222222222222e-06, "loss": -0.0449, "num_tokens": 33914851.0, "reward": 1.8240714073181152, "reward_std": 0.3135382831096649, "rewards/accuracy_reward_step": 0.625, "rewards/final_brier_reward_step": 0.7405382394790649, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8369972109794617, "step": 120 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.514165890027959, "calib/avg_num_step_conf": 6.7109375, "calib/ece": 0.12331054687499991, "calib/final_conf_rate": 1.0, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.0234375, "calib/gap": 0.01594529356943164, "calib/mean_conf": 0.623720703125, "calib/mu_c": 0.6306344827586208, "calib/mu_w": 0.6146891891891891, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.98828125, "calib/pce": 0.09031249999999995, "calib/std_conf": 0.15473907322430916, "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": 1358.0, "completions/max_terminated_length": 1358.0, "completions/mean_length": 530.14453125, "completions/mean_terminated_length": 532.2235717773438, "completions/min_length": 0.0, "completions/min_terminated_length": 140.0, "epoch": 0.12906666666666666, "grad_norm": 0.00623322231695056, "learning_rate": 2.1944444444444445e-06, "loss": 0.0287, "num_tokens": 34209128.0, "reward": 1.729459524154663, "reward_std": 0.42823922634124756, "rewards/accuracy_reward_step": 0.56640625, "rewards/final_brier_reward_step": 0.7250615358352661, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8177761435508728, "step": 121 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5791911814431595, "calib/avg_num_step_conf": 6.64453125, "calib/ece": 0.07603174603174608, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.01984126984126984, "calib/gap": 0.03973625481842147, "calib/mean_conf": 0.6505555555555556, "calib/mu_c": 0.6652201257861634, "calib/mu_w": 0.6254838709677419, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.04781746031746032, "calib/std_conf": 0.11764761450745415, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2006.0, "completions/max_terminated_length": 2006.0, "completions/mean_length": 509.4609375, "completions/mean_terminated_length": 511.4588623046875, "completions/min_length": 0.0, "completions/min_terminated_length": 165.0, "epoch": 0.13013333333333332, "grad_norm": 0.005936076398938894, "learning_rate": 2.166666666666667e-06, "loss": -0.0152, "num_tokens": 34500398.0, "reward": 1.8247861862182617, "reward_std": 0.30627113580703735, "rewards/accuracy_reward_step": 0.62109375, "rewards/final_brier_reward_step": 0.7593749761581421, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8444574475288391, "step": 122 }, { "calib/answer_extract_rate": 0.96484375, "calib/auroc": 0.44736315086782374, "calib/avg_num_step_conf": 6.90234375, "calib/ece": 0.17028744939271262, "calib/final_conf_rate": 0.96484375, "calib/format_rate": 0.96484375, "calib/frac_conf_gt_0.9": 0.024291497975708502, "calib/gap": -0.022474966622162973, "calib/mean_conf": 0.6587004048582996, "calib/mu_c": 0.6489642857142857, "calib/mu_w": 0.6714392523364486, "calib/nonempty_final_conf_rate": 0.96484375, "calib/nonempty_reasoning_rate": 0.9921875, "calib/nonempty_step_conf_rate": 0.9921875, "calib/pce": 0.13109311740890692, "calib/std_conf": 0.14508163037931543, "calib/step_conf_rate": 0.9921875, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 1719.0, "completions/max_terminated_length": 1719.0, "completions/mean_length": 577.81640625, "completions/mean_terminated_length": 586.9881591796875, "completions/min_length": 0.0, "completions/min_terminated_length": 222.0, "epoch": 0.1312, "grad_norm": 0.006853913888335228, "learning_rate": 2.138888888888889e-06, "loss": -0.0255, "num_tokens": 34807111.0, "reward": 1.6731445789337158, "reward_std": 0.503294050693512, "rewards/accuracy_reward_step": 0.546875, "rewards/final_brier_reward_step": 0.6888322830200195, "rewards/format_reward_step": 0.96484375, "rewards/stepwise_brier_reward": 0.7928085327148438, "step": 123 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5398951247165533, "calib/avg_num_step_conf": 6.99609375, "calib/ece": 0.06869047619047626, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.031746031746031744, "calib/gap": 0.020119047619047592, "calib/mean_conf": 0.6659126984126983, "calib/mu_c": 0.6726190476190477, "calib/mu_w": 0.6525000000000001, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.03396825396825398, "calib/std_conf": 0.12300616321126623, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/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": 549.3828125, "completions/mean_terminated_length": 551.5372924804688, "completions/min_length": 0.0, "completions/min_terminated_length": 173.0, "epoch": 0.13226666666666667, "grad_norm": 0.006010471843183041, "learning_rate": 2.1111111111111114e-06, "loss": 0.0056, "num_tokens": 35108073.0, "reward": 1.875612497329712, "reward_std": 0.3735159635543823, "rewards/accuracy_reward_step": 0.65625, "rewards/final_brier_reward_step": 0.7595324516296387, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8366676568984985, "step": 124 }, { "calib/answer_extract_rate": 0.95703125, "calib/auroc": 0.43928271812080544, "calib/avg_num_step_conf": 7.4609375, "calib/ece": 0.15224489795918367, "calib/final_conf_rate": 0.95703125, "calib/format_rate": 0.953125, "calib/frac_conf_gt_0.9": 0.024489795918367346, "calib/gap": -0.029948965324384846, "calib/mean_conf": 0.671265306122449, "calib/mu_c": 0.6595302013422818, "calib/mu_w": 0.6894791666666666, "calib/nonempty_final_conf_rate": 0.95703125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.10767346938775507, "calib/std_conf": 0.12023629380247575, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2412.0, "completions/max_terminated_length": 2412.0, "completions/mean_length": 595.88671875, "completions/mean_terminated_length": 605.3452758789062, "completions/min_length": 0.0, "completions/min_terminated_length": 200.0, "epoch": 0.13333333333333333, "grad_norm": 0.005993053782731295, "learning_rate": 2.0833333333333334e-06, "loss": 0.0266, "num_tokens": 35418932.0, "reward": 1.7210444211959839, "reward_std": 0.37963348627090454, "rewards/accuracy_reward_step": 0.58203125, "rewards/final_brier_reward_step": 0.6943820118904114, "rewards/format_reward_step": 0.953125, "rewards/stepwise_brier_reward": 0.791358232498169, "step": 125 }, { "calib/answer_extract_rate": 0.921875, "calib/auroc": 0.5743065900546215, "calib/avg_num_step_conf": 8.41015625, "calib/ece": 0.19197478991596642, "calib/final_conf_rate": 0.9296875, "calib/format_rate": 0.921875, "calib/frac_conf_gt_0.9": 0.025210084033613446, "calib/gap": 0.03010214939348821, "calib/mean_conf": 0.705252100840336, "calib/mu_c": 0.7192913385826772, "calib/mu_w": 0.689189189189189, "calib/nonempty_final_conf_rate": 0.9296875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.18180672268907566, "calib/std_conf": 0.12255225045761145, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2774.0, "completions/max_terminated_length": 2774.0, "completions/mean_length": 603.53125, "completions/mean_terminated_length": 610.687744140625, "completions/min_length": 0.0, "completions/min_terminated_length": 165.0, "epoch": 0.1344, "grad_norm": 0.00570314098149538, "learning_rate": 2.0555555555555555e-06, "loss": -0.0133, "num_tokens": 35732404.0, "reward": 1.5615825653076172, "reward_std": 0.512492299079895, "rewards/accuracy_reward_step": 0.49609375, "rewards/final_brier_reward_step": 0.6676222681999207, "rewards/format_reward_step": 0.921875, "rewards/stepwise_brier_reward": 0.7583954334259033, "step": 126 }, { "calib/answer_extract_rate": 0.90234375, "calib/auroc": 0.46693657219973006, "calib/avg_num_step_conf": 8.6640625, "calib/ece": 0.24180086580086577, "calib/final_conf_rate": 0.90234375, "calib/format_rate": 0.90234375, "calib/frac_conf_gt_0.9": 0.025974025974025976, "calib/gap": -0.018092217723796655, "calib/mean_conf": 0.7110995670995671, "calib/mu_c": 0.7021709401709402, "calib/mu_w": 0.7202631578947368, "calib/nonempty_final_conf_rate": 0.90234375, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.22320346320346315, "calib/std_conf": 0.12446659716299954, "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": 2630.0, "completions/max_terminated_length": 2630.0, "completions/mean_length": 604.640625, "completions/mean_terminated_length": 604.640625, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.13546666666666668, "grad_norm": 0.006331843789666891, "learning_rate": 2.027777777777778e-06, "loss": 0.0164, "num_tokens": 36044368.0, "reward": 1.4707010984420776, "reward_std": 0.5850914716720581, "rewards/accuracy_reward_step": 0.45703125, "rewards/final_brier_reward_step": 0.6168801188468933, "rewards/format_reward_step": 0.90234375, "rewards/stepwise_brier_reward": 0.7190492153167725, "step": 127 }, { "calib/answer_extract_rate": 0.9296875, "calib/auroc": 0.5129772629772629, "calib/avg_num_step_conf": 8.52734375, "calib/ece": 0.18708860759493673, "calib/final_conf_rate": 0.92578125, "calib/format_rate": 0.91796875, "calib/frac_conf_gt_0.9": 0.046413502109704644, "calib/gap": 0.00798155298155323, "calib/mean_conf": 0.7025316455696202, "calib/mu_c": 0.7062698412698415, "calib/mu_w": 0.6982882882882883, "calib/nonempty_final_conf_rate": 0.92578125, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.98828125, "calib/pce": 0.17898734177215192, "calib/std_conf": 0.1290550043440286, "calib/step_conf_rate": 0.98828125, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2703.0, "completions/max_terminated_length": 2703.0, "completions/mean_length": 580.26953125, "completions/mean_terminated_length": 589.4801635742188, "completions/min_length": 0.0, "completions/min_terminated_length": 196.0, "epoch": 0.13653333333333334, "grad_norm": 0.006742393597960472, "learning_rate": 2.0000000000000003e-06, "loss": 0.0181, "num_tokens": 36353085.0, "reward": 1.5568528175354004, "reward_std": 0.5391932725906372, "rewards/accuracy_reward_step": 0.49609375, "rewards/final_brier_reward_step": 0.6504441499710083, "rewards/format_reward_step": 0.91796875, "rewards/stepwise_brier_reward": 0.764467179775238, "step": 128 }, { "calib/answer_extract_rate": 0.94921875, "calib/auroc": 0.48491863075196406, "calib/avg_num_step_conf": 8.38671875, "calib/ece": 0.1300008230452675, "calib/final_conf_rate": 0.94921875, "calib/format_rate": 0.94921875, "calib/frac_conf_gt_0.9": 0.07407407407407407, "calib/gap": -0.005633333333333601, "calib/mean_conf": 0.7064617283950618, "calib/mu_c": 0.7041666666666666, "calib/mu_w": 0.7098000000000002, "calib/nonempty_final_conf_rate": 0.94921875, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.12193497942386831, "calib/std_conf": 0.1351615073237689, "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": 1506.0, "completions/max_terminated_length": 1506.0, "completions/mean_length": 536.23828125, "completions/mean_terminated_length": 538.3411865234375, "completions/min_length": 0.0, "completions/min_terminated_length": 190.0, "epoch": 0.1376, "grad_norm": 0.008170709945261478, "learning_rate": 1.9722222222222224e-06, "loss": 0.018, "num_tokens": 36646250.0, "reward": 1.6871366500854492, "reward_std": 0.48693281412124634, "rewards/accuracy_reward_step": 0.5625, "rewards/final_brier_reward_step": 0.6878214478492737, "rewards/format_reward_step": 0.94921875, "rewards/stepwise_brier_reward": 0.7872874736785889, "step": 129 }, { "calib/answer_extract_rate": 0.96875, "calib/auroc": 0.5813823857302117, "calib/avg_num_step_conf": 8.31640625, "calib/ece": 0.11141129032258065, "calib/final_conf_rate": 0.96875, "calib/format_rate": 0.96875, "calib/frac_conf_gt_0.9": 0.016129032258064516, "calib/gap": 0.05376532887402452, "calib/mean_conf": 0.6972983870967743, "calib/mu_c": 0.7172435897435897, "calib/mu_w": 0.6634782608695652, "calib/nonempty_final_conf_rate": 0.96875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.08983870967741939, "calib/std_conf": 0.12928603446264397, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1355.0, "completions/max_terminated_length": 1355.0, "completions/mean_length": 530.015625, "completions/mean_terminated_length": 532.0941772460938, "completions/min_length": 0.0, "completions/min_terminated_length": 176.0, "epoch": 0.13866666666666666, "grad_norm": 0.005932880565524101, "learning_rate": 1.944444444444445e-06, "loss": 0.0132, "num_tokens": 36940726.0, "reward": 1.787769079208374, "reward_std": 0.3610919713973999, "rewards/accuracy_reward_step": 0.609375, "rewards/final_brier_reward_step": 0.7462925910949707, "rewards/format_reward_step": 0.96875, "rewards/stepwise_brier_reward": 0.8110337257385254, "step": 130 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5192367844854551, "calib/avg_num_step_conf": 7.80078125, "calib/ece": 0.2592086614173228, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.031496062992125984, "calib/gap": 0.01488576790741325, "calib/mean_conf": 0.6967755905511811, "calib/mu_c": 0.7049217391304348, "calib/mu_w": 0.6900359712230215, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.25161417322834645, "calib/std_conf": 0.13428815252328039, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1466.0, "completions/max_terminated_length": 1466.0, "completions/mean_length": 501.70703125, "completions/mean_terminated_length": 503.6745300292969, "completions/min_length": 0.0, "completions/min_terminated_length": 209.0, "epoch": 0.13973333333333332, "grad_norm": 0.006347838789224625, "learning_rate": 1.916666666666667e-06, "loss": 0.0122, "num_tokens": 37228875.0, "reward": 1.5378695726394653, "reward_std": 0.3169251084327698, "rewards/accuracy_reward_step": 0.44921875, "rewards/final_brier_reward_step": 0.6767011880874634, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.7950894832611084, "step": 131 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.515625, "calib/avg_num_step_conf": 8.390625, "calib/ece": 0.13456126482213437, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.07114624505928854, "calib/gap": 0.02521303763440852, "calib/mean_conf": 0.711794466403162, "calib/mu_c": 0.7210624999999999, "calib/mu_w": 0.6958494623655914, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.10697233201581027, "calib/std_conf": 0.14592896252549273, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2496.0, "completions/max_terminated_length": 2496.0, "completions/mean_length": 581.73046875, "completions/mean_terminated_length": 581.73046875, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.1408, "grad_norm": 0.005685857962816954, "learning_rate": 1.888888888888889e-06, "loss": 0.0326, "num_tokens": 37536894.0, "reward": 1.8229343891143799, "reward_std": 0.472642719745636, "rewards/accuracy_reward_step": 0.625, "rewards/final_brier_reward_step": 0.742849588394165, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8223249912261963, "step": 132 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.5128384687208217, "calib/avg_num_step_conf": 9.17578125, "calib/ece": 0.3725290836653388, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.10358565737051793, "calib/gap": -0.011041710017340223, "calib/mean_conf": 0.7242836653386454, "calib/mu_c": 0.7175530612244898, "calib/mu_w": 0.72859477124183, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.35318725099601606, "calib/std_conf": 0.16828970992382916, "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": 2732.0, "completions/max_terminated_length": 2732.0, "completions/mean_length": 662.41015625, "completions/mean_terminated_length": 667.6259765625, "completions/min_length": 0.0, "completions/min_terminated_length": 208.0, "epoch": 0.14186666666666667, "grad_norm": 0.005634021945297718, "learning_rate": 1.8611111111111113e-06, "loss": 0.0653, "num_tokens": 37866319.0, "reward": 1.404353141784668, "reward_std": 0.4524964392185211, "rewards/accuracy_reward_step": 0.3828125, "rewards/final_brier_reward_step": 0.6049235463142395, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.754676342010498, "step": 133 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.507124842370744, "calib/avg_num_step_conf": 8.765625, "calib/ece": 0.24087301587301585, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.07539682539682539, "calib/gap": 0.020315258511979795, "calib/mean_conf": 0.7121428571428571, "calib/mu_c": 0.7226229508196721, "calib/mu_w": 0.7023076923076923, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.23444444444444446, "calib/std_conf": 0.1489172943086789, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2216.0, "completions/max_terminated_length": 2216.0, "completions/mean_length": 660.40234375, "completions/mean_terminated_length": 660.40234375, "completions/min_length": 245.0, "completions/min_terminated_length": 245.0, "epoch": 0.14293333333333333, "grad_norm": 0.005534619092941284, "learning_rate": 1.8333333333333333e-06, "loss": 0.05, "num_tokens": 38197838.0, "reward": 1.5717178583145142, "reward_std": 0.49151453375816345, "rewards/accuracy_reward_step": 0.4765625, "rewards/final_brier_reward_step": 0.6755093336105347, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.783237099647522, "step": 134 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5043229497774953, "calib/avg_num_step_conf": 9.3125, "calib/ece": 0.1841343873517787, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.06324110671936758, "calib/gap": -0.00010489510489486431, "calib/mean_conf": 0.7147588932806325, "calib/mu_c": 0.7147132867132868, "calib/mu_w": 0.7148181818181817, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.16683794466403168, "calib/std_conf": 0.14822907563496882, "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": 2640.0, "completions/max_terminated_length": 2640.0, "completions/mean_length": 618.63671875, "completions/mean_terminated_length": 621.0628051757812, "completions/min_length": 0.0, "completions/min_terminated_length": 201.0, "epoch": 0.144, "grad_norm": 0.006134727504104376, "learning_rate": 1.8055555555555557e-06, "loss": 0.0462, "num_tokens": 38515593.0, "reward": 1.705806851387024, "reward_std": 0.5274053812026978, "rewards/accuracy_reward_step": 0.55859375, "rewards/final_brier_reward_step": 0.6979269981384277, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8049877882003784, "step": 135 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5596153846153846, "calib/avg_num_step_conf": 8.90625, "calib/ece": 0.24175390625, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.05078125, "calib/gap": 0.02924481074481078, "calib/mean_conf": 0.7178554687500001, "calib/mu_c": 0.7327063492063491, "calib/mu_w": 0.7034615384615384, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.23371093749999997, "calib/std_conf": 0.14617925174924032, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1580.0, "completions/max_terminated_length": 1580.0, "completions/mean_length": 551.07421875, "completions/mean_terminated_length": 553.2353515625, "completions/min_length": 0.0, "completions/min_terminated_length": 178.0, "epoch": 0.14506666666666668, "grad_norm": 0.006309314630925655, "learning_rate": 1.777777777777778e-06, "loss": 0.0197, "num_tokens": 38818660.0, "reward": 1.6099199056625366, "reward_std": 0.4166410565376282, "rewards/accuracy_reward_step": 0.4921875, "rewards/final_brier_reward_step": 0.6923854351043701, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.7941691875457764, "step": 136 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.520057215511761, "calib/avg_num_step_conf": 8.80078125, "calib/ece": 0.22968127490039836, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.055776892430278883, "calib/gap": 0.01902542911633831, "calib/mean_conf": 0.720597609561753, "calib/mu_c": 0.7297692307692308, "calib/mu_w": 0.7107438016528925, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.2161752988047808, "calib/std_conf": 0.11636271014810072, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2340.0, "completions/max_terminated_length": 2340.0, "completions/mean_length": 591.52734375, "completions/mean_terminated_length": 596.18505859375, "completions/min_length": 0.0, "completions/min_terminated_length": 294.0, "epoch": 0.14613333333333334, "grad_norm": 0.005561066325753927, "learning_rate": 1.75e-06, "loss": -0.0109, "num_tokens": 39130579.0, "reward": 1.6218444108963013, "reward_std": 0.41467946767807007, "rewards/accuracy_reward_step": 0.5078125, "rewards/final_brier_reward_step": 0.6914331912994385, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.7881319522857666, "step": 137 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.43930884102021933, "calib/avg_num_step_conf": 8.47265625, "calib/ece": 0.14973725490196083, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.058823529411764705, "calib/gap": -0.02187676754328005, "calib/mean_conf": 0.7270705882352942, "calib/mu_c": 0.7190062111801242, "calib/mu_w": 0.7408829787234043, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.12271764705882357, "calib/std_conf": 0.1250286356127403, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1293.0, "completions/max_terminated_length": 1293.0, "completions/mean_length": 560.51953125, "completions/mean_terminated_length": 562.7176513671875, "completions/min_length": 0.0, "completions/min_terminated_length": 164.0, "epoch": 0.1472, "grad_norm": 0.0072199031710624695, "learning_rate": 1.7222222222222224e-06, "loss": 0.008, "num_tokens": 39431912.0, "reward": 1.8308939933776855, "reward_std": 0.3743973970413208, "rewards/accuracy_reward_step": 0.62890625, "rewards/final_brier_reward_step": 0.7294247150421143, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.828526496887207, "step": 138 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5791357370095439, "calib/avg_num_step_conf": 7.87109375, "calib/ece": 0.06840624999999995, "calib/final_conf_rate": 1.0, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.0234375, "calib/gap": 0.039593849416755034, "calib/mean_conf": 0.702234375, "calib/mu_c": 0.7164634146341463, "calib/mu_w": 0.6768695652173913, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.06500781249999996, "calib/std_conf": 0.11701805979573998, "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": 1511.0, "completions/max_terminated_length": 1511.0, "completions/mean_length": 525.1171875, "completions/mean_terminated_length": 527.176513671875, "completions/min_length": 0.0, "completions/min_terminated_length": 199.0, "epoch": 0.14826666666666666, "grad_norm": 0.006623912137001753, "learning_rate": 1.6944444444444446e-06, "loss": 0.011, "num_tokens": 39722942.0, "reward": 1.8607146739959717, "reward_std": 0.3195941746234894, "rewards/accuracy_reward_step": 0.640625, "rewards/final_brier_reward_step": 0.768987774848938, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8379333019256592, "step": 139 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5192708333333333, "calib/avg_num_step_conf": 8.38671875, "calib/ece": 0.17385703125000002, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.0859375, "calib/gap": -0.0011012499999999426, "calib/mean_conf": 0.7141117187499999, "calib/mu_c": 0.71369875, "calib/mu_w": 0.7148, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.131484375, "calib/std_conf": 0.14477266506551192, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1459.0, "completions/max_terminated_length": 1459.0, "completions/mean_length": 551.7890625, "completions/mean_terminated_length": 553.9530029296875, "completions/min_length": 0.0, "completions/min_terminated_length": 169.0, "epoch": 0.14933333333333335, "grad_norm": 0.0066830990836024284, "learning_rate": 1.6666666666666667e-06, "loss": 0.0158, "num_tokens": 40022720.0, "reward": 1.8299082517623901, "reward_std": 0.37425488233566284, "rewards/accuracy_reward_step": 0.625, "rewards/final_brier_reward_step": 0.7362087965011597, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.8334243297576904, "step": 140 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5309180852611669, "calib/avg_num_step_conf": 8.1875, "calib/ece": 0.08554545454545451, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.043478260869565216, "calib/gap": 0.01825356467335948, "calib/mean_conf": 0.6929525691699605, "calib/mu_c": 0.6986522988505747, "calib/mu_w": 0.6803987341772152, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.04537549407114623, "calib/std_conf": 0.1527392431775241, "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": 2198.0, "completions/max_terminated_length": 2198.0, "completions/mean_length": 613.75390625, "completions/mean_terminated_length": 613.75390625, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.1504, "grad_norm": 0.005835548508912325, "learning_rate": 1.638888888888889e-06, "loss": 0.0212, "num_tokens": 40340441.0, "reward": 1.9075510501861572, "reward_std": 0.3685436248779297, "rewards/accuracy_reward_step": 0.6796875, "rewards/final_brier_reward_step": 0.7568938732147217, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8264350295066833, "step": 141 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.45533906882591085, "calib/avg_num_step_conf": 8.76171875, "calib/ece": 0.18972265624999998, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.05859375, "calib/gap": -0.018941801619433263, "calib/mean_conf": 0.71433984375, "calib/mu_c": 0.7066447368421053, "calib/mu_w": 0.7255865384615385, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.15515625, "calib/std_conf": 0.14971691299741519, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1450.0, "completions/max_terminated_length": 1450.0, "completions/mean_length": 596.15625, "completions/mean_terminated_length": 598.494140625, "completions/min_length": 0.0, "completions/min_terminated_length": 206.0, "epoch": 0.15146666666666667, "grad_norm": 0.0058297100476920605, "learning_rate": 1.6111111111111113e-06, "loss": 0.0106, "num_tokens": 40651721.0, "reward": 1.769550085067749, "reward_std": 0.3588702082633972, "rewards/accuracy_reward_step": 0.59375, "rewards/final_brier_reward_step": 0.7126940488815308, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.8030059337615967, "step": 142 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.4417295558579735, "calib/avg_num_step_conf": 8.36328125, "calib/ece": 0.21302431372549016, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.058823529411764705, "calib/gap": -0.016315452183595203, "calib/mean_conf": 0.7023090196078432, "calib/mu_c": 0.6947591240875911, "calib/mu_w": 0.7110745762711863, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.18903921568627444, "calib/std_conf": 0.14765126758889496, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2764.0, "completions/max_terminated_length": 2764.0, "completions/mean_length": 567.76171875, "completions/mean_terminated_length": 567.76171875, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.15253333333333333, "grad_norm": 0.006419370416551828, "learning_rate": 1.5833333333333333e-06, "loss": 0.0226, "num_tokens": 40957908.0, "reward": 1.6711039543151855, "reward_std": 0.28329381346702576, "rewards/accuracy_reward_step": 0.53515625, "rewards/final_brier_reward_step": 0.6915198564529419, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.7897710204124451, "step": 143 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5479624542124543, "calib/avg_num_step_conf": 8.2578125, "calib/ece": 0.08637795275590551, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.07086614173228346, "calib/gap": 0.027614468864468944, "calib/mean_conf": 0.7293700787401575, "calib/mu_c": 0.7371978021978023, "calib/mu_w": 0.7095833333333333, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.049606299212598404, "calib/std_conf": 0.11912573937602051, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1690.0, "completions/max_terminated_length": 1690.0, "completions/mean_length": 553.3984375, "completions/mean_terminated_length": 555.5686645507812, "completions/min_length": 0.0, "completions/min_terminated_length": 133.0, "epoch": 0.1536, "grad_norm": 0.0074084168300032616, "learning_rate": 1.5555555555555558e-06, "loss": 0.0199, "num_tokens": 41257210.0, "reward": 1.9680758714675903, "reward_std": 0.30323338508605957, "rewards/accuracy_reward_step": 0.7109375, "rewards/final_brier_reward_step": 0.7875484228134155, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.8347547054290771, "step": 144 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.4422513051562611, "calib/avg_num_step_conf": 8.44140625, "calib/ece": 0.18599414062500003, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.0859375, "calib/gap": -0.022122970750196758, "calib/mean_conf": 0.7263496093749999, "calib/mu_c": 0.7195225988700565, "calib/mu_w": 0.7416455696202533, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.11046875000000003, "calib/std_conf": 0.1572008438665245, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1301.0, "completions/max_terminated_length": 1301.0, "completions/mean_length": 544.40234375, "completions/mean_terminated_length": 546.5372924804688, "completions/min_length": 0.0, "completions/min_terminated_length": 180.0, "epoch": 0.15466666666666667, "grad_norm": 0.006819000467658043, "learning_rate": 1.527777777777778e-06, "loss": 0.0309, "num_tokens": 41552785.0, "reward": 1.9274626970291138, "reward_std": 0.3706474304199219, "rewards/accuracy_reward_step": 0.69140625, "rewards/final_brier_reward_step": 0.7512627243995667, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.810150682926178, "step": 145 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.52141065830721, "calib/avg_num_step_conf": 8.37890625, "calib/ece": 0.3054117647058824, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.07450980392156863, "calib/gap": 0.01604702194357366, "calib/mean_conf": 0.7367843137254901, "calib/mu_c": 0.7459090909090909, "calib/mu_w": 0.7298620689655172, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.3054117647058824, "calib/std_conf": 0.11195172519237315, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2351.0, "completions/max_terminated_length": 2351.0, "completions/mean_length": 574.33203125, "completions/mean_terminated_length": 574.33203125, "completions/min_length": 224.0, "completions/min_terminated_length": 224.0, "epoch": 0.15573333333333333, "grad_norm": 0.006699086166918278, "learning_rate": 1.5e-06, "loss": 0.0462, "num_tokens": 41860534.0, "reward": 1.5063364505767822, "reward_std": 0.3869948983192444, "rewards/accuracy_reward_step": 0.4296875, "rewards/final_brier_reward_step": 0.6542070508003235, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8008264303207397, "step": 146 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5055374796646227, "calib/avg_num_step_conf": 8.71484375, "calib/ece": 0.25288063241106723, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.11462450592885376, "calib/gap": -0.0037813540232762843, "calib/mean_conf": 0.750992885375494, "calib/mu_c": 0.7491694656488549, "calib/mu_w": 0.7529508196721312, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.2430434782608696, "calib/std_conf": 0.13074846906492196, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2149.0, "completions/max_terminated_length": 2149.0, "completions/mean_length": 581.87890625, "completions/mean_terminated_length": 586.4606323242188, "completions/min_length": 0.0, "completions/min_terminated_length": 226.0, "epoch": 0.1568, "grad_norm": 0.006568745709955692, "learning_rate": 1.4722222222222225e-06, "loss": 0.0068, "num_tokens": 42166679.0, "reward": 1.6215184926986694, "reward_std": 0.43788471817970276, "rewards/accuracy_reward_step": 0.51171875, "rewards/final_brier_reward_step": 0.6690147519111633, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.7701841592788696, "step": 147 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5240274599542334, "calib/avg_num_step_conf": 8.1484375, "calib/ece": 0.14878906250000007, "calib/final_conf_rate": 1.0, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.09765625, "calib/gap": 0.00248512585812366, "calib/mean_conf": 0.7577734375, "calib/mu_c": 0.758695652173913, "calib/mu_w": 0.7562105263157893, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.13882812500000008, "calib/std_conf": 0.11943275532463277, "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": 1660.0, "completions/max_terminated_length": 1660.0, "completions/mean_length": 549.9140625, "completions/mean_terminated_length": 552.0706176757812, "completions/min_length": 0.0, "completions/min_terminated_length": 165.0, "epoch": 0.15786666666666666, "grad_norm": 0.006827679928392172, "learning_rate": 1.4444444444444445e-06, "loss": 0.0382, "num_tokens": 42466073.0, "reward": 1.820509910583496, "reward_std": 0.3338512182235718, "rewards/accuracy_reward_step": 0.62890625, "rewards/final_brier_reward_step": 0.7346500158309937, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.7817646265029907, "step": 148 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.48321791752448684, "calib/avg_num_step_conf": 8.82421875, "calib/ece": 0.20283464566929135, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.13385826771653545, "calib/gap": 0.0025928005490049033, "calib/mean_conf": 0.7376377952755905, "calib/mu_c": 0.7388321167883212, "calib/mu_w": 0.7362393162393163, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.20055118110236222, "calib/std_conf": 0.14199343831414551, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/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": 613.84375, "completions/mean_terminated_length": 613.84375, "completions/min_length": 116.0, "completions/min_terminated_length": 116.0, "epoch": 0.15893333333333334, "grad_norm": 0.00622599758207798, "learning_rate": 1.4166666666666667e-06, "loss": 0.0498, "num_tokens": 42781177.0, "reward": 1.6633987426757812, "reward_std": 0.4651468098163605, "rewards/accuracy_reward_step": 0.53515625, "rewards/final_brier_reward_step": 0.6879491806030273, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.7703335285186768, "step": 149 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.49910181561557704, "calib/avg_num_step_conf": 7.53515625, "calib/ece": 0.19075396825396823, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.07936507936507936, "calib/gap": 0.0009296208378777315, "calib/mean_conf": 0.7274999999999999, "calib/mu_c": 0.7279020979020979, "calib/mu_w": 0.7269724770642202, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.17539682539682538, "calib/std_conf": 0.13409788239749376, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2547.0, "completions/max_terminated_length": 2547.0, "completions/mean_length": 524.2890625, "completions/mean_terminated_length": 526.3451538085938, "completions/min_length": 0.0, "completions/min_terminated_length": 192.0, "epoch": 0.16, "grad_norm": 0.007395992986857891, "learning_rate": 1.3888888888888892e-06, "loss": 0.0096, "num_tokens": 43073859.0, "reward": 1.7042057514190674, "reward_std": 0.3107081353664398, "rewards/accuracy_reward_step": 0.55859375, "rewards/final_brier_reward_step": 0.7002965211868286, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.7962145805358887, "step": 150 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.555656363408835, "calib/avg_num_step_conf": 8.640625, "calib/ece": 0.263808300395257, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.08300395256916997, "calib/gap": 0.03188971968464516, "calib/mean_conf": 0.7280494071146245, "calib/mu_c": 0.7445614754098361, "calib/mu_w": 0.7126717557251909, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.25482213438735185, "calib/std_conf": 0.16013216666255328, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2394.0, "completions/max_terminated_length": 2394.0, "completions/mean_length": 607.421875, "completions/mean_terminated_length": 609.803955078125, "completions/min_length": 0.0, "completions/min_terminated_length": 267.0, "epoch": 0.16106666666666666, "grad_norm": 0.006049179006367922, "learning_rate": 1.3611111111111112e-06, "loss": 0.0174, "num_tokens": 43389887.0, "reward": 1.573657751083374, "reward_std": 0.43454253673553467, "rewards/accuracy_reward_step": 0.4765625, "rewards/final_brier_reward_step": 0.6721926927566528, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.7865009307861328, "step": 151 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.45484375000000005, "calib/avg_num_step_conf": 8.60546875, "calib/ece": 0.25730711462450595, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.07509881422924901, "calib/gap": -0.01792099374999989, "calib/mean_conf": 0.7207956521739131, "calib/mu_c": 0.71194140625, "calib/mu_w": 0.7298623999999999, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.23608695652173917, "calib/std_conf": 0.17156045875656895, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2029.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 591.3515625, "completions/mean_terminated_length": 596.0078735351562, "completions/min_length": 0.0, "completions/min_terminated_length": 242.0, "epoch": 0.16213333333333332, "grad_norm": 0.006287906318902969, "learning_rate": 1.3333333333333334e-06, "loss": 0.0157, "num_tokens": 43700169.0, "reward": 1.6060374975204468, "reward_std": 0.4876149892807007, "rewards/accuracy_reward_step": 0.5, "rewards/final_brier_reward_step": 0.657676637172699, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.7899107933044434, "step": 152 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.4231292517006802, "calib/avg_num_step_conf": 8.3125, "calib/ece": 0.24569246031746023, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.11507936507936507, "calib/gap": -0.027608843537414884, "calib/mean_conf": 0.7682757936507936, "calib/mu_c": 0.7567721088435374, "calib/mu_w": 0.7843809523809523, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.21531746031746024, "calib/std_conf": 0.13920559880784064, "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": 1869.0, "completions/max_terminated_length": 1869.0, "completions/mean_length": 592.48828125, "completions/mean_terminated_length": 597.153564453125, "completions/min_length": 0.0, "completions/min_terminated_length": 188.0, "epoch": 0.1632, "grad_norm": 0.00608451571315527, "learning_rate": 1.3055555555555556e-06, "loss": 0.033, "num_tokens": 44012670.0, "reward": 1.7243762016296387, "reward_std": 0.3136652112007141, "rewards/accuracy_reward_step": 0.578125, "rewards/final_brier_reward_step": 0.6791612505912781, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.7808442115783691, "step": 153 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.6319457179534201, "calib/avg_num_step_conf": 8.3046875, "calib/ece": 0.297578125, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.1015625, "calib/gap": 0.04381808178984037, "calib/mean_conf": 0.7460156249999998, "calib/mu_c": 0.768780487804878, "calib/mu_w": 0.7249624060150376, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.28156250000000005, "calib/std_conf": 0.1485468141895321, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1381.0, "completions/max_terminated_length": 1381.0, "completions/mean_length": 566.55859375, "completions/mean_terminated_length": 568.7803955078125, "completions/min_length": 0.0, "completions/min_terminated_length": 216.0, "epoch": 0.16426666666666667, "grad_norm": 0.006061206571757793, "learning_rate": 1.2777777777777779e-06, "loss": -0.0173, "num_tokens": 44315653.0, "reward": 1.5989031791687012, "reward_std": 0.36871886253356934, "rewards/accuracy_reward_step": 0.48046875, "rewards/final_brier_reward_step": 0.6796757578849792, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.8331245183944702, "step": 154 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.6213455772113943, "calib/avg_num_step_conf": 7.828125, "calib/ece": 0.18095275590551194, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.03937007874015748, "calib/gap": 0.05698175912043968, "calib/mean_conf": 0.7225275590551181, "calib/mu_c": 0.7485507246376811, "calib/mu_w": 0.6915689655172415, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.18008661417322847, "calib/std_conf": 0.12500679205305654, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2152.0, "completions/max_terminated_length": 2152.0, "completions/mean_length": 521.68359375, "completions/mean_terminated_length": 523.7294311523438, "completions/min_length": 0.0, "completions/min_terminated_length": 217.0, "epoch": 0.16533333333333333, "grad_norm": 0.006489680614322424, "learning_rate": 1.25e-06, "loss": 0.0256, "num_tokens": 44609924.0, "reward": 1.6893417835235596, "reward_std": 0.337790310382843, "rewards/accuracy_reward_step": 0.5390625, "rewards/final_brier_reward_step": 0.7266840934753418, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.8119329214096069, "step": 155 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5695497246517655, "calib/avg_num_step_conf": 8.46875, "calib/ece": 0.1824503968253968, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.0992063492063492, "calib/gap": 0.04562244897959178, "calib/mean_conf": 0.7469940476190476, "calib/mu_c": 0.7660034013605441, "calib/mu_w": 0.7203809523809523, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.9921875, "calib/pce": 0.17305555555555552, "calib/std_conf": 0.15827228448467448, "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": 2744.0, "completions/max_terminated_length": 2744.0, "completions/mean_length": 636.39453125, "completions/mean_terminated_length": 636.39453125, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "epoch": 0.1664, "grad_norm": 0.0057357740588486195, "learning_rate": 1.2222222222222223e-06, "loss": 0.0022, "num_tokens": 44931105.0, "reward": 1.7241895198822021, "reward_std": 0.40883588790893555, "rewards/accuracy_reward_step": 0.57421875, "rewards/final_brier_reward_step": 0.7088333368301392, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.7894869446754456, "step": 156 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.6025133584009499, "calib/avg_num_step_conf": 8.66796875, "calib/ece": 0.13722656249999998, "calib/final_conf_rate": 1.0, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.08984375, "calib/gap": 0.04760406359258518, "calib/mean_conf": 0.7624609375, "calib/mu_c": 0.7797546012269938, "calib/mu_w": 0.7321505376344086, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.131484375, "calib/std_conf": 0.13436990701649346, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1338.0, "completions/max_terminated_length": 1338.0, "completions/mean_length": 587.13671875, "completions/mean_terminated_length": 589.4392700195312, "completions/min_length": 0.0, "completions/min_terminated_length": 199.0, "epoch": 0.16746666666666668, "grad_norm": 0.00636444753035903, "learning_rate": 1.1944444444444446e-06, "loss": 0.0053, "num_tokens": 45238644.0, "reward": 1.8468554019927979, "reward_std": 0.4306687116622925, "rewards/accuracy_reward_step": 0.63671875, "rewards/final_brier_reward_step": 0.7564671635627747, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8184540271759033, "step": 157 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.46067561434221616, "calib/avg_num_step_conf": 8.8515625, "calib/ece": 0.17631835937499998, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.10546875, "calib/gap": -0.004201355118978234, "calib/mean_conf": 0.7459472656249999, "calib/mu_c": 0.7443553459119496, "calib/mu_w": 0.7485567010309279, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.1505859375, "calib/std_conf": 0.16237854280784206, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1552.0, "completions/max_terminated_length": 1552.0, "completions/mean_length": 581.41015625, "completions/mean_terminated_length": 583.6902465820312, "completions/min_length": 0.0, "completions/min_terminated_length": 176.0, "epoch": 0.16853333333333334, "grad_norm": 0.005845364648848772, "learning_rate": 1.1666666666666668e-06, "loss": -0.0096, "num_tokens": 45546229.0, "reward": 1.8103464841842651, "reward_std": 0.4110521674156189, "rewards/accuracy_reward_step": 0.62109375, "rewards/final_brier_reward_step": 0.7207310199737549, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.7940922975540161, "step": 158 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5165912518853695, "calib/avg_num_step_conf": 7.79296875, "calib/ece": 0.20928853754940718, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.09881422924901186, "calib/gap": 0.007342885872297478, "calib/mean_conf": 0.7245454545454546, "calib/mu_c": 0.7279411764705882, "calib/mu_w": 0.7205982905982907, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.98828125, "calib/pce": 0.19814229249011864, "calib/std_conf": 0.14070506713716469, "calib/step_conf_rate": 0.98828125, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.01171875, "completions/max_length": 2290.0, "completions/max_terminated_length": 2290.0, "completions/mean_length": 559.28515625, "completions/mean_terminated_length": 565.9169921875, "completions/min_length": 0.0, "completions/min_terminated_length": 249.0, "epoch": 0.1696, "grad_norm": 0.006066169124096632, "learning_rate": 1.138888888888889e-06, "loss": 0.0084, "num_tokens": 45847694.0, "reward": 1.6596370935440063, "reward_std": 0.35814833641052246, "rewards/accuracy_reward_step": 0.53125, "rewards/final_brier_reward_step": 0.6832605600357056, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.8146629333496094, "step": 159 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.5246598639455782, "calib/avg_num_step_conf": 8.13671875, "calib/ece": 0.19099601593625493, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.09561752988047809, "calib/gap": 0.019807692307692415, "calib/mean_conf": 0.7517928286852589, "calib/mu_c": 0.76, "calib/mu_w": 0.7401923076923076, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.1785657370517928, "calib/std_conf": 0.1320945759687546, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2924.0, "completions/max_terminated_length": 2924.0, "completions/mean_length": 596.421875, "completions/mean_terminated_length": 605.888916015625, "completions/min_length": 0.0, "completions/min_terminated_length": 246.0, "epoch": 0.17066666666666666, "grad_norm": 0.006081533618271351, "learning_rate": 1.111111111111111e-06, "loss": -0.0276, "num_tokens": 46158722.0, "reward": 1.729783296585083, "reward_std": 0.4630153775215149, "rewards/accuracy_reward_step": 0.57421875, "rewards/final_brier_reward_step": 0.7078007459640503, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.8050822019577026, "step": 160 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5335677972003621, "calib/avg_num_step_conf": 7.4765625, "calib/ece": 0.09707031249999994, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.078125, "calib/gap": 0.01529563340065454, "calib/mean_conf": 0.7466015625, "calib/mu_c": 0.7515606936416184, "calib/mu_w": 0.7362650602409638, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.08394531249999997, "calib/std_conf": 0.12594243723844079, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1449.0, "completions/max_terminated_length": 1449.0, "completions/mean_length": 543.171875, "completions/mean_terminated_length": 545.302001953125, "completions/min_length": 0.0, "completions/min_terminated_length": 247.0, "epoch": 0.17173333333333332, "grad_norm": 0.0063605401664972305, "learning_rate": 1.0833333333333335e-06, "loss": 0.0464, "num_tokens": 46455198.0, "reward": 1.9115158319473267, "reward_std": 0.31776556372642517, "rewards/accuracy_reward_step": 0.67578125, "rewards/final_brier_reward_step": 0.7667245864868164, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.8246514797210693, "step": 161 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5099041308332158, "calib/avg_num_step_conf": 7.80078125, "calib/ece": 0.18270588235294116, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.13725490196078433, "calib/gap": 0.002021006626250932, "calib/mean_conf": 0.7380784313725489, "calib/mu_c": 0.7387283236994218, "calib/mu_w": 0.7367073170731708, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.12117647058823527, "calib/std_conf": 0.18241687262188017, "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": 1460.0, "completions/max_terminated_length": 1460.0, "completions/mean_length": 554.24609375, "completions/mean_terminated_length": 556.4196166992188, "completions/min_length": 0.0, "completions/min_terminated_length": 211.0, "epoch": 0.1728, "grad_norm": 0.006147413980215788, "learning_rate": 1.0555555555555557e-06, "loss": 0.0195, "num_tokens": 46754733.0, "reward": 1.9030340909957886, "reward_std": 0.35743248462677, "rewards/accuracy_reward_step": 0.67578125, "rewards/final_brier_reward_step": 0.7392222881317139, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.8338513970375061, "step": 162 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.6477477477477479, "calib/avg_num_step_conf": 8.44921875, "calib/ece": 0.21186852589641442, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.1553784860557769, "calib/gap": 0.09593314028314026, "calib/mean_conf": 0.7517968127490041, "calib/mu_c": 0.7942214285714286, "calib/mu_w": 0.6982882882882884, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.2029482071713148, "calib/std_conf": 0.1860408497044822, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2675.0, "completions/max_terminated_length": 2675.0, "completions/mean_length": 640.5546875, "completions/mean_terminated_length": 643.0667114257812, "completions/min_length": 0.0, "completions/min_terminated_length": 200.0, "epoch": 0.17386666666666667, "grad_norm": 0.005643155891448259, "learning_rate": 1.0277777777777777e-06, "loss": 0.0272, "num_tokens": 47077051.0, "reward": 1.6887856721878052, "reward_std": 0.40363791584968567, "rewards/accuracy_reward_step": 0.546875, "rewards/final_brier_reward_step": 0.7141788005828857, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.7987761497497559, "step": 163 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5760030864197531, "calib/avg_num_step_conf": 8.5234375, "calib/ece": 0.23340079365079366, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.15873015873015872, "calib/gap": 0.03497453703703712, "calib/mean_conf": 0.7790595238095237, "calib/mu_c": 0.7940486111111111, "calib/mu_w": 0.759074074074074, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.22051587301587303, "calib/std_conf": 0.16092937062553864, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2838.0, "completions/max_terminated_length": 2838.0, "completions/mean_length": 651.28515625, "completions/mean_terminated_length": 653.8392333984375, "completions/min_length": 0.0, "completions/min_terminated_length": 298.0, "epoch": 0.17493333333333333, "grad_norm": 0.005333068314939737, "learning_rate": 1.0000000000000002e-06, "loss": 0.0115, "num_tokens": 47403420.0, "reward": 1.7017207145690918, "reward_std": 0.43776828050613403, "rewards/accuracy_reward_step": 0.5625, "rewards/final_brier_reward_step": 0.6922357082366943, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.7708970308303833, "step": 164 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.49162531017369726, "calib/avg_num_step_conf": 7.890625, "calib/ece": 0.2785826771653544, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.12598425196850394, "calib/gap": -0.009270471464019936, "calib/mean_conf": 0.7594488188976377, "calib/mu_c": 0.7549230769230768, "calib/mu_w": 0.7641935483870967, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.26311023622047247, "calib/std_conf": 0.14451552016815059, "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": 1500.0, "completions/max_terminated_length": 1500.0, "completions/mean_length": 607.41796875, "completions/mean_terminated_length": 609.800048828125, "completions/min_length": 0.0, "completions/min_terminated_length": 212.0, "epoch": 0.176, "grad_norm": 0.005638706963509321, "learning_rate": 9.722222222222224e-07, "loss": -0.0105, "num_tokens": 47717999.0, "reward": 1.6242128610610962, "reward_std": 0.31716349720954895, "rewards/accuracy_reward_step": 0.5078125, "rewards/final_brier_reward_step": 0.6581156253814697, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.8074854612350464, "step": 165 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5904713804713804, "calib/avg_num_step_conf": 8.44140625, "calib/ece": 0.1903843137254902, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.17254901960784313, "calib/gap": 0.02168888888888887, "calib/mean_conf": 0.7611450980392157, "calib/mu_c": 0.7687999999999999, "calib/mu_w": 0.7471111111111111, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.15223529411764702, "calib/std_conf": 0.1795983496078211, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2516.0, "completions/max_terminated_length": 2516.0, "completions/mean_length": 643.046875, "completions/mean_terminated_length": 643.046875, "completions/min_length": 269.0, "completions/min_terminated_length": 269.0, "epoch": 0.17706666666666668, "grad_norm": 0.005603920668363571, "learning_rate": 9.444444444444445e-07, "loss": 0.0247, "num_tokens": 48042307.0, "reward": 1.851743459701538, "reward_std": 0.3167978525161743, "rewards/accuracy_reward_step": 0.64453125, "rewards/final_brier_reward_step": 0.7333853244781494, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8142135143280029, "step": 166 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5299003322259137, "calib/avg_num_step_conf": 8.58984375, "calib/ece": 0.17288281250000007, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.19140625, "calib/gap": 0.009408084163898045, "calib/mean_conf": 0.7772734374999999, "calib/mu_c": 0.7803604651162791, "calib/mu_w": 0.7709523809523811, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.13914062500000007, "calib/std_conf": 0.16700510914919217, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1658.0, "completions/max_terminated_length": 1658.0, "completions/mean_length": 603.44921875, "completions/mean_terminated_length": 605.8157348632812, "completions/min_length": 0.0, "completions/min_terminated_length": 260.0, "epoch": 0.17813333333333334, "grad_norm": 0.005436644423753023, "learning_rate": 9.166666666666666e-07, "loss": 0.0147, "num_tokens": 48355902.0, "reward": 1.8921669721603394, "reward_std": 0.3189052939414978, "rewards/accuracy_reward_step": 0.671875, "rewards/final_brier_reward_step": 0.7446897029876709, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.7927284240722656, "step": 167 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.6103058510638298, "calib/avg_num_step_conf": 7.9609375, "calib/ece": 0.19651574803149618, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.21653543307086615, "calib/gap": 0.038319148936170055, "calib/mean_conf": 0.7913188976377953, "calib/mu_c": 0.8055, "calib/mu_w": 0.7671808510638299, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.17895669291338595, "calib/std_conf": 0.13917194895223228, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1558.0, "completions/max_terminated_length": 1558.0, "completions/mean_length": 619.37890625, "completions/mean_terminated_length": 621.807861328125, "completions/min_length": 0.0, "completions/min_terminated_length": 280.0, "epoch": 0.1792, "grad_norm": 0.00561784952878952, "learning_rate": 8.88888888888889e-07, "loss": 0.0245, "num_tokens": 48672639.0, "reward": 1.8194490671157837, "reward_std": 0.3768016993999481, "rewards/accuracy_reward_step": 0.625, "rewards/final_brier_reward_step": 0.7335515022277832, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.809869647026062, "step": 168 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.6004505362015355, "calib/avg_num_step_conf": 7.8203125, "calib/ece": 0.22264531249999994, "calib/final_conf_rate": 1.0, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.20703125, "calib/gap": 0.03448876197728323, "calib/mean_conf": 0.7799328125, "calib/mu_c": 0.7938091503267974, "calib/mu_w": 0.7593203883495142, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.20246093749999997, "calib/std_conf": 0.15693214950525544, "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": 1598.0, "completions/max_terminated_length": 1598.0, "completions/mean_length": 598.87109375, "completions/mean_terminated_length": 601.2196655273438, "completions/min_length": 0.0, "completions/min_terminated_length": 238.0, "epoch": 0.18026666666666666, "grad_norm": 0.005926609039306641, "learning_rate": 8.611111111111112e-07, "loss": 0.0079, "num_tokens": 48983638.0, "reward": 1.7784779071807861, "reward_std": 0.32226821780204773, "rewards/accuracy_reward_step": 0.59765625, "rewards/final_brier_reward_step": 0.7149896025657654, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8207975625991821, "step": 169 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.5239749587231701, "calib/avg_num_step_conf": 7.671875, "calib/ece": 0.21224000000000004, "calib/final_conf_rate": 0.9765625, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.164, "calib/gap": -0.00724545954870659, "calib/mean_conf": 0.7721600000000001, "calib/mu_c": 0.7694936708860759, "calib/mu_w": 0.7767391304347825, "calib/nonempty_final_conf_rate": 0.9765625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.17620000000000002, "calib/std_conf": 0.15087787909431918, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2999.0, "completions/max_terminated_length": 2999.0, "completions/mean_length": 631.6796875, "completions/mean_terminated_length": 634.1569213867188, "completions/min_length": 0.0, "completions/min_terminated_length": 215.0, "epoch": 0.18133333333333335, "grad_norm": 0.00526054622605443, "learning_rate": 8.333333333333333e-07, "loss": 0.054, "num_tokens": 49303004.0, "reward": 1.7908600568771362, "reward_std": 0.3574887812137604, "rewards/accuracy_reward_step": 0.6171875, "rewards/final_brier_reward_step": 0.7047312259674072, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.8024589419364929, "step": 170 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.6458461728090719, "calib/avg_num_step_conf": 7.7109375, "calib/ece": 0.254, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.12941176470588237, "calib/gap": 0.057351781092074505, "calib/mean_conf": 0.7755686274509803, "calib/mu_c": 0.8030075187969925, "calib/mu_w": 0.745655737704918, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.254, "calib/std_conf": 0.10994994492182425, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1436.0, "completions/max_terminated_length": 1436.0, "completions/mean_length": 563.43359375, "completions/mean_terminated_length": 565.6431884765625, "completions/min_length": 0.0, "completions/min_terminated_length": 238.0, "epoch": 0.1824, "grad_norm": 0.006021480076014996, "learning_rate": 8.055555555555557e-07, "loss": 0.0186, "num_tokens": 49607643.0, "reward": 1.6539881229400635, "reward_std": 0.40474262833595276, "rewards/accuracy_reward_step": 0.51953125, "rewards/final_brier_reward_step": 0.699738621711731, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8068389892578125, "step": 171 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.5774999999999999, "calib/avg_num_step_conf": 7.4609375, "calib/ece": 0.18006299212598423, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.14566929133858267, "calib/gap": 0.03285324675324708, "calib/mean_conf": 0.7748188976377953, "calib/mu_c": 0.7877532467532469, "calib/mu_w": 0.7548999999999998, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 0.9921875, "calib/nonempty_step_conf_rate": 0.984375, "calib/pce": 0.17429133858267717, "calib/std_conf": 0.12417662156648765, "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": 1644.0, "completions/max_terminated_length": 1644.0, "completions/mean_length": 580.1171875, "completions/mean_terminated_length": 582.3922119140625, "completions/min_length": 0.0, "completions/min_terminated_length": 259.0, "epoch": 0.18346666666666667, "grad_norm": 0.006254405248910189, "learning_rate": 7.777777777777779e-07, "loss": -0.0029, "num_tokens": 49913009.0, "reward": 1.7794864177703857, "reward_std": 0.38404881954193115, "rewards/accuracy_reward_step": 0.6015625, "rewards/final_brier_reward_step": 0.720995306968689, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.8188252449035645, "step": 172 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.6240579521829522, "calib/avg_num_step_conf": 8.125, "calib/ece": 0.2017857142857143, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.1388888888888889, "calib/gap": 0.04792359667359669, "calib/mean_conf": 0.7803571428571429, "calib/mu_c": 0.8001351351351352, "calib/mu_w": 0.7522115384615385, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.1974206349206349, "calib/std_conf": 0.12449370254269537, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2505.0, "completions/max_terminated_length": 2505.0, "completions/mean_length": 644.03515625, "completions/mean_terminated_length": 644.03515625, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.18453333333333333, "grad_norm": 0.005519158206880093, "learning_rate": 7.5e-07, "loss": 0.0041, "num_tokens": 50234546.0, "reward": 1.7319469451904297, "reward_std": 0.4004300832748413, "rewards/accuracy_reward_step": 0.578125, "rewards/final_brier_reward_step": 0.7167074084281921, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.7735806703567505, "step": 173 }, { "calib/answer_extract_rate": 0.96875, "calib/auroc": 0.5729740361919748, "calib/avg_num_step_conf": 8.5234375, "calib/ece": 0.2819838056680162, "calib/final_conf_rate": 0.96484375, "calib/format_rate": 0.96484375, "calib/frac_conf_gt_0.9": 0.11740890688259109, "calib/gap": 0.026395226855494336, "calib/mean_conf": 0.770242914979757, "calib/mu_c": 0.7833870967741936, "calib/mu_w": 0.7569918699186993, "calib/nonempty_final_conf_rate": 0.96484375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.2751012145748988, "calib/std_conf": 0.12553666859668175, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2669.0, "completions/max_terminated_length": 2669.0, "completions/mean_length": 693.6875, "completions/mean_terminated_length": 696.4078979492188, "completions/min_length": 0.0, "completions/min_terminated_length": 298.0, "epoch": 0.1856, "grad_norm": 0.005586102604866028, "learning_rate": 7.222222222222222e-07, "loss": 0.067, "num_tokens": 50569866.0, "reward": 1.5690968036651611, "reward_std": 0.5186349153518677, "rewards/accuracy_reward_step": 0.48828125, "rewards/final_brier_reward_step": 0.651752769947052, "rewards/format_reward_step": 0.96484375, "rewards/stepwise_brier_reward": 0.7652595043182373, "step": 174 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.5354398516163222, "calib/avg_num_step_conf": 7.88671875, "calib/ece": 0.3674, "calib/final_conf_rate": 0.9765625, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.176, "calib/gap": 0.03294250132485421, "calib/mean_conf": 0.7753999999999999, "calib/mu_c": 0.7949019607843135, "calib/mu_w": 0.7619594594594593, "calib/nonempty_final_conf_rate": 0.9765625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.3674, "calib/std_conf": 0.14140169730240157, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2415.0, "completions/max_terminated_length": 2415.0, "completions/mean_length": 614.84375, "completions/mean_terminated_length": 622.1343994140625, "completions/min_length": 0.0, "completions/min_terminated_length": 270.0, "epoch": 0.18666666666666668, "grad_norm": 0.005461685825139284, "learning_rate": 6.944444444444446e-07, "loss": -0.006, "num_tokens": 50886594.0, "reward": 1.4294147491455078, "reward_std": 0.39421916007995605, "rewards/accuracy_reward_step": 0.3984375, "rewards/final_brier_reward_step": 0.6048831939697266, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.7690262198448181, "step": 175 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.6574375000000001, "calib/avg_num_step_conf": 7.796875, "calib/ece": 0.2786916996047431, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.21739130434782608, "calib/gap": 0.07296593749999991, "calib/mean_conf": 0.7740355731225297, "calib/mu_c": 0.8100859375, "calib/mu_w": 0.7371200000000001, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.27339920948616603, "calib/std_conf": 0.15209681595056485, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1631.0, "completions/max_terminated_length": 1631.0, "completions/mean_length": 596.72265625, "completions/mean_terminated_length": 599.0628051757812, "completions/min_length": 0.0, "completions/min_terminated_length": 256.0, "epoch": 0.18773333333333334, "grad_norm": 0.005617086309939623, "learning_rate": 6.666666666666667e-07, "loss": -0.0241, "num_tokens": 51196923.0, "reward": 1.6214114427566528, "reward_std": 0.4137636423110962, "rewards/accuracy_reward_step": 0.5, "rewards/final_brier_reward_step": 0.6833947896957397, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.8256884813308716, "step": 176 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.530801209372638, "calib/avg_num_step_conf": 8.4375, "calib/ece": 0.21896862745098036, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.20784313725490197, "calib/gap": 0.024574452003023217, "calib/mean_conf": 0.7754627450980391, "calib/mu_c": 0.7858707482993197, "calib/mu_w": 0.7612962962962965, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.2089803921568627, "calib/std_conf": 0.1508815844748487, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2498.0, "completions/max_terminated_length": 2498.0, "completions/mean_length": 649.16015625, "completions/mean_terminated_length": 649.16015625, "completions/min_length": 272.0, "completions/min_terminated_length": 272.0, "epoch": 0.1888, "grad_norm": 0.005589707288891077, "learning_rate": 6.388888888888889e-07, "loss": 0.0288, "num_tokens": 51520444.0, "reward": 1.7286344766616821, "reward_std": 0.3408302366733551, "rewards/accuracy_reward_step": 0.57421875, "rewards/final_brier_reward_step": 0.7027286291122437, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.7743090391159058, "step": 177 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.49151768236982907, "calib/avg_num_step_conf": 8.05078125, "calib/ece": 0.17182352941176476, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.10980392156862745, "calib/gap": 0.002634085867153857, "calib/mean_conf": 0.774137254901961, "calib/mu_c": 0.7751392405063292, "calib/mu_w": 0.7725051546391754, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.1631764705882353, "calib/std_conf": 0.12293706657295218, "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": 1582.0, "completions/max_terminated_length": 1582.0, "completions/mean_length": 604.2734375, "completions/mean_terminated_length": 606.6431884765625, "completions/min_length": 0.0, "completions/min_terminated_length": 182.0, "epoch": 0.18986666666666666, "grad_norm": 0.005940349772572517, "learning_rate": 6.111111111111112e-07, "loss": 0.0218, "num_tokens": 51834714.0, "reward": 1.8063156604766846, "reward_std": 0.38152995705604553, "rewards/accuracy_reward_step": 0.6171875, "rewards/final_brier_reward_step": 0.7199667692184448, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.81779545545578, "step": 178 }, { "calib/answer_extract_rate": 0.98046875, "calib/auroc": 0.5555480743334231, "calib/avg_num_step_conf": 8.05859375, "calib/ece": 0.14297619047619042, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.9765625, "calib/frac_conf_gt_0.9": 0.12301587301587301, "calib/gap": 0.02721653649340139, "calib/mean_conf": 0.7644047619047619, "calib/mu_c": 0.7745569620253164, "calib/mu_w": 0.747340425531915, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.14019841269841263, "calib/std_conf": 0.11459747902788475, "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": 2732.0, "completions/max_terminated_length": 2732.0, "completions/mean_length": 634.21484375, "completions/mean_terminated_length": 636.7020263671875, "completions/min_length": 0.0, "completions/min_terminated_length": 273.0, "epoch": 0.19093333333333334, "grad_norm": 0.005728751886636019, "learning_rate": 5.833333333333334e-07, "loss": 0.0347, "num_tokens": 52156841.0, "reward": 1.7972197532653809, "reward_std": 0.5223114490509033, "rewards/accuracy_reward_step": 0.6171875, "rewards/final_brier_reward_step": 0.7318102121353149, "rewards/format_reward_step": 0.9765625, "rewards/stepwise_brier_reward": 0.8008187413215637, "step": 179 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5480562238824871, "calib/avg_num_step_conf": 8.20703125, "calib/ece": 0.2318039215686274, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.22745098039215686, "calib/gap": 0.021933012536406227, "calib/mean_conf": 0.8161176470588235, "calib/mu_c": 0.8252348993288591, "calib/mu_w": 0.8033018867924528, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.2318039215686274, "calib/std_conf": 0.10088736625055425, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1434.0, "completions/max_terminated_length": 1434.0, "completions/mean_length": 666.36328125, "completions/mean_terminated_length": 668.9765014648438, "completions/min_length": 0.0, "completions/min_terminated_length": 284.0, "epoch": 0.192, "grad_norm": 0.005887932609766722, "learning_rate": 5.555555555555555e-07, "loss": -0.0157, "num_tokens": 52484790.0, "reward": 1.7513511180877686, "reward_std": 0.3672969937324524, "rewards/accuracy_reward_step": 0.58203125, "rewards/final_brier_reward_step": 0.7011027336120605, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8199268579483032, "step": 180 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.6483791313156895, "calib/avg_num_step_conf": 7.59765625, "calib/ece": 0.19843137254901963, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.14901960784313725, "calib/gap": 0.059138913511459945, "calib/mean_conf": 0.7617254901960785, "calib/mu_c": 0.7863087248322147, "calib/mu_w": 0.7271698113207548, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.187921568627451, "calib/std_conf": 0.12648863832281423, "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": 2143.0, "completions/max_terminated_length": 2143.0, "completions/mean_length": 581.5234375, "completions/mean_terminated_length": 581.5234375, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.19306666666666666, "grad_norm": 0.0064398800022900105, "learning_rate": 5.277777777777779e-07, "loss": 0.0379, "num_tokens": 52793428.0, "reward": 1.7602028846740723, "reward_std": 0.4153677821159363, "rewards/accuracy_reward_step": 0.58203125, "rewards/final_brier_reward_step": 0.7335976362228394, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.8306518197059631, "step": 181 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5894155844155844, "calib/avg_num_step_conf": 8.55859375, "calib/ece": 0.19767716535433064, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.19291338582677164, "calib/gap": 0.04086493506493505, "calib/mean_conf": 0.8039763779527558, "calib/mu_c": 0.820064935064935, "calib/mu_w": 0.7792, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.19767716535433064, "calib/std_conf": 0.10442726923610096, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2965.0, "completions/max_terminated_length": 2965.0, "completions/mean_length": 619.0390625, "completions/mean_terminated_length": 621.4666748046875, "completions/min_length": 0.0, "completions/min_terminated_length": 310.0, "epoch": 0.19413333333333332, "grad_norm": 0.006538864690810442, "learning_rate": 5.000000000000001e-07, "loss": 0.0273, "num_tokens": 53111566.0, "reward": 1.7762978076934814, "reward_std": 0.3718116283416748, "rewards/accuracy_reward_step": 0.6015625, "rewards/final_brier_reward_step": 0.7240335941314697, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.7952203750610352, "step": 182 }, { "calib/answer_extract_rate": 0.9921875, "calib/auroc": 0.4991611259553843, "calib/avg_num_step_conf": 8.1171875, "calib/ece": 0.2506929133858268, "calib/final_conf_rate": 0.9921875, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.1377952755905512, "calib/gap": 0.013665071770334936, "calib/mean_conf": 0.7677007874015748, "calib/mu_c": 0.7742105263157895, "calib/mu_w": 0.7605454545454545, "calib/nonempty_final_conf_rate": 0.9921875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.24738582677165355, "calib/std_conf": 0.1415837480642303, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 2721.0, "completions/max_terminated_length": 2721.0, "completions/mean_length": 659.27734375, "completions/mean_terminated_length": 659.27734375, "completions/min_length": 316.0, "completions/min_terminated_length": 316.0, "epoch": 0.1952, "grad_norm": 0.006364805158227682, "learning_rate": 4.7222222222222226e-07, "loss": 0.0394, "num_tokens": 53440525.0, "reward": 1.6405270099639893, "reward_std": 0.4377848505973816, "rewards/accuracy_reward_step": 0.51953125, "rewards/final_brier_reward_step": 0.6724599003791809, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.7880856990814209, "step": 183 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.6033903054076772, "calib/avg_num_step_conf": 8.1953125, "calib/ece": 0.13531746031746036, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.09523809523809523, "calib/gap": 0.03188007845334817, "calib/mean_conf": 0.7913492063492064, "calib/mu_c": 0.8022289156626505, "calib/mu_w": 0.7703488372093024, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.13396825396825401, "calib/std_conf": 0.08124871842313493, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2980.0, "completions/max_terminated_length": 2980.0, "completions/mean_length": 628.3359375, "completions/mean_terminated_length": 630.800048828125, "completions/min_length": 0.0, "completions/min_terminated_length": 235.0, "epoch": 0.19626666666666667, "grad_norm": 0.006428469903767109, "learning_rate": 4.444444444444445e-07, "loss": 0.005, "num_tokens": 53760163.0, "reward": 1.8640644550323486, "reward_std": 0.4095544219017029, "rewards/accuracy_reward_step": 0.65234375, "rewards/final_brier_reward_step": 0.7518515586853027, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.8294061422348022, "step": 184 }, { "calib/answer_extract_rate": 0.97265625, "calib/auroc": 0.5972053872053872, "calib/avg_num_step_conf": 8.36328125, "calib/ece": 0.17036144578313245, "calib/final_conf_rate": 0.97265625, "calib/format_rate": 0.97265625, "calib/frac_conf_gt_0.9": 0.12048192771084337, "calib/gap": 0.04870303030303047, "calib/mean_conf": 0.7723694779116466, "calib/mu_c": 0.7917333333333334, "calib/mu_w": 0.7430303030303029, "calib/nonempty_final_conf_rate": 0.97265625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.17016064257028107, "calib/std_conf": 0.11749223691405654, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2951.0, "completions/max_terminated_length": 2951.0, "completions/mean_length": 624.37109375, "completions/mean_terminated_length": 634.2817993164062, "completions/min_length": 0.0, "completions/min_terminated_length": 285.0, "epoch": 0.19733333333333333, "grad_norm": 0.005404021590948105, "learning_rate": 4.1666666666666667e-07, "loss": 0.0102, "num_tokens": 54080426.0, "reward": 1.743666172027588, "reward_std": 0.42095649242401123, "rewards/accuracy_reward_step": 0.5859375, "rewards/final_brier_reward_step": 0.7208616733551025, "rewards/format_reward_step": 0.97265625, "rewards/stepwise_brier_reward": 0.7928655743598938, "step": 185 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.5332303931030046, "calib/avg_num_step_conf": 8.359375, "calib/ece": 0.21289062499999992, "calib/final_conf_rate": 1.0, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.16796875, "calib/gap": -0.005581933989577292, "calib/mean_conf": 0.79203125, "calib/mu_c": 0.7898726114649681, "calib/mu_w": 0.7954545454545454, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.1958203124999999, "calib/std_conf": 0.13073625271300038, "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": 1422.0, "completions/max_terminated_length": 1422.0, "completions/mean_length": 616.8046875, "completions/mean_terminated_length": 619.2235717773438, "completions/min_length": 0.0, "completions/min_terminated_length": 240.0, "epoch": 0.1984, "grad_norm": 0.0060129775665700436, "learning_rate": 3.8888888888888895e-07, "loss": 0.0066, "num_tokens": 54396872.0, "reward": 1.8011512756347656, "reward_std": 0.4412759244441986, "rewards/accuracy_reward_step": 0.61328125, "rewards/final_brier_reward_step": 0.7091492414474487, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8235810399055481, "step": 186 }, { "calib/answer_extract_rate": 0.97265625, "calib/auroc": 0.5164765100671141, "calib/avg_num_step_conf": 8.43359375, "calib/ece": 0.2052208835341367, "calib/final_conf_rate": 0.97265625, "calib/format_rate": 0.97265625, "calib/frac_conf_gt_0.9": 0.18473895582329317, "calib/gap": 0.0023261744966442954, "calib/mean_conf": 0.7779919678714858, "calib/mu_c": 0.7789261744966444, "calib/mu_w": 0.7766000000000001, "calib/nonempty_final_conf_rate": 0.97265625, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.19240963855421703, "calib/std_conf": 0.12697588525973325, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2578.0, "completions/max_terminated_length": 2578.0, "completions/mean_length": 662.85546875, "completions/mean_terminated_length": 665.4549560546875, "completions/min_length": 0.0, "completions/min_terminated_length": 235.0, "epoch": 0.19946666666666665, "grad_norm": 0.006127472966909409, "learning_rate": 3.611111111111111e-07, "loss": 0.0553, "num_tokens": 54721611.0, "reward": 1.7329193353652954, "reward_std": 0.46366044878959656, "rewards/accuracy_reward_step": 0.58203125, "rewards/final_brier_reward_step": 0.6929405927658081, "rewards/format_reward_step": 0.97265625, "rewards/stepwise_brier_reward": 0.801236629486084, "step": 187 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5703052728954672, "calib/avg_num_step_conf": 8.671875, "calib/ece": 0.18329411764705883, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.18823529411764706, "calib/gap": 0.02609092110479727, "calib/mean_conf": 0.8009411764705883, "calib/mu_c": 0.8105590062111802, "calib/mu_w": 0.7844680851063829, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 0.99609375, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.17643137254901964, "calib/std_conf": 0.12097275852231473, "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": 674.6484375, "completions/mean_terminated_length": 677.2941284179688, "completions/min_length": 0.0, "completions/min_terminated_length": 229.0, "epoch": 0.20053333333333334, "grad_norm": 0.005286317318677902, "learning_rate": 3.3333333333333335e-07, "loss": 0.0347, "num_tokens": 55051897.0, "reward": 1.8274012804031372, "reward_std": 0.3512701988220215, "rewards/accuracy_reward_step": 0.62890625, "rewards/final_brier_reward_step": 0.7331405878067017, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8108394145965576, "step": 188 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.6542526101349632, "calib/avg_num_step_conf": 7.7578125, "calib/ece": 0.20054687499999996, "calib/final_conf_rate": 1.0, "calib/format_rate": 1.0, "calib/frac_conf_gt_0.9": 0.1953125, "calib/gap": 0.05696205755029282, "calib/mean_conf": 0.802109375, "calib/mu_c": 0.824805194805195, "calib/mu_w": 0.7678431372549022, "calib/nonempty_final_conf_rate": 1.0, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.20054687499999996, "calib/std_conf": 0.10354009144823745, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1169.0, "completions/max_terminated_length": 1169.0, "completions/mean_length": 602.08203125, "completions/mean_terminated_length": 604.4431762695312, "completions/min_length": 0.0, "completions/min_terminated_length": 307.0, "epoch": 0.2016, "grad_norm": 0.006643631961196661, "learning_rate": 3.055555555555556e-07, "loss": 0.0229, "num_tokens": 55367302.0, "reward": 1.7953031063079834, "reward_std": 0.3850705027580261, "rewards/accuracy_reward_step": 0.6015625, "rewards/final_brier_reward_step": 0.7366812229156494, "rewards/format_reward_step": 1.0, "rewards/stepwise_brier_reward": 0.8351558446884155, "step": 189 }, { "calib/answer_extract_rate": 0.98828125, "calib/auroc": 0.5420884889388827, "calib/avg_num_step_conf": 8.19140625, "calib/ece": 0.3062055335968379, "calib/final_conf_rate": 0.98828125, "calib/format_rate": 0.98828125, "calib/frac_conf_gt_0.9": 0.16996047430830039, "calib/gap": 0.022463442069741713, "calib/mean_conf": 0.7915810276679842, "calib/mu_c": 0.8028571428571429, "calib/mu_w": 0.7803937007874012, "calib/nonempty_final_conf_rate": 0.98828125, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.29988142292490116, "calib/std_conf": 0.1337800623956951, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2218.0, "completions/max_terminated_length": 2218.0, "completions/mean_length": 684.3046875, "completions/mean_terminated_length": 686.98828125, "completions/min_length": 0.0, "completions/min_terminated_length": 319.0, "epoch": 0.20266666666666666, "grad_norm": 0.0055880011059343815, "learning_rate": 2.7777777777777776e-07, "loss": 0.0032, "num_tokens": 55701596.0, "reward": 1.5926530361175537, "reward_std": 0.4001859128475189, "rewards/accuracy_reward_step": 0.4921875, "rewards/final_brier_reward_step": 0.6494613885879517, "rewards/format_reward_step": 0.98828125, "rewards/stepwise_brier_reward": 0.7914630174636841, "step": 190 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.612427316590375, "calib/avg_num_step_conf": 8.90625, "calib/ece": 0.2735294117647059, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.18823529411764706, "calib/gap": 0.04240381046641106, "calib/mean_conf": 0.8017647058823529, "calib/mu_c": 0.8213868613138686, "calib/mu_w": 0.7789830508474576, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.2690196078431373, "calib/std_conf": 0.12408576860386342, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1740.0, "completions/max_terminated_length": 1740.0, "completions/mean_length": 633.66015625, "completions/mean_terminated_length": 636.1451416015625, "completions/min_length": 0.0, "completions/min_terminated_length": 227.0, "epoch": 0.20373333333333332, "grad_norm": 0.0062250238843262196, "learning_rate": 2.5000000000000004e-07, "loss": -0.007, "num_tokens": 56021485.0, "reward": 1.6655709743499756, "reward_std": 0.38111838698387146, "rewards/accuracy_reward_step": 0.53515625, "rewards/final_brier_reward_step": 0.684425413608551, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.7747330665588379, "step": 191 }, { "calib/answer_extract_rate": 0.9765625, "calib/auroc": 0.5559180772709652, "calib/avg_num_step_conf": 7.99609375, "calib/ece": 0.228207171314741, "calib/final_conf_rate": 0.98046875, "calib/format_rate": 0.97265625, "calib/frac_conf_gt_0.9": 0.20318725099601595, "calib/gap": 0.04281044062540418, "calib/mean_conf": 0.7939442231075697, "calib/mu_c": 0.8125352112676058, "calib/mu_w": 0.7697247706422016, "calib/nonempty_final_conf_rate": 0.98046875, "calib/nonempty_reasoning_rate": 0.9921875, "calib/nonempty_step_conf_rate": 0.98828125, "calib/pce": 0.228207171314741, "calib/std_conf": 0.1371478862603115, "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": 2987.0, "completions/max_terminated_length": 2987.0, "completions/mean_length": 664.9765625, "completions/mean_terminated_length": 667.5843505859375, "completions/min_length": 0.0, "completions/min_terminated_length": 201.0, "epoch": 0.2048, "grad_norm": 0.006628439761698246, "learning_rate": 2.2222222222222224e-07, "loss": 0.053, "num_tokens": 56350199.0, "reward": 1.6849875450134277, "reward_std": 0.4626771807670593, "rewards/accuracy_reward_step": 0.5546875, "rewards/final_brier_reward_step": 0.6843519806861877, "rewards/format_reward_step": 0.97265625, "rewards/stepwise_brier_reward": 0.7821609377861023, "step": 192 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.4930168513540096, "calib/avg_num_step_conf": 8.15234375, "calib/ece": 0.199563492063492, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.19444444444444445, "calib/gap": 0.018912202478526208, "calib/mean_conf": 0.7928174603174603, "calib/mu_c": 0.8003973509933775, "calib/mu_w": 0.7814851485148513, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.19658730158730148, "calib/std_conf": 0.13440786664392912, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2690.0, "completions/max_terminated_length": 2690.0, "completions/mean_length": 646.6171875, "completions/mean_terminated_length": 649.1529541015625, "completions/min_length": 0.0, "completions/min_terminated_length": 298.0, "epoch": 0.20586666666666667, "grad_norm": 0.005542590282857418, "learning_rate": 1.9444444444444447e-07, "loss": 0.0745, "num_tokens": 56674949.0, "reward": 1.7428584098815918, "reward_std": 0.5023571848869324, "rewards/accuracy_reward_step": 0.58984375, "rewards/final_brier_reward_step": 0.6989473104476929, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.7724858522415161, "step": 193 }, { "calib/answer_extract_rate": 1.0, "calib/auroc": 0.6438970492655661, "calib/avg_num_step_conf": 8.0703125, "calib/ece": 0.2206274509803921, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.17254901960784313, "calib/gap": 0.03630573248407665, "calib/mean_conf": 0.7823529411764705, "calib/mu_c": 0.7963057324840765, "calib/mu_w": 0.7599999999999999, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.19364705882352937, "calib/std_conf": 0.13939066606292952, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1364.0, "completions/max_terminated_length": 1364.0, "completions/mean_length": 586.16796875, "completions/mean_terminated_length": 588.4666748046875, "completions/min_length": 0.0, "completions/min_terminated_length": 299.0, "epoch": 0.20693333333333333, "grad_norm": 0.006334380712360144, "learning_rate": 1.6666666666666668e-07, "loss": 0.0042, "num_tokens": 56984456.0, "reward": 1.8093762397766113, "reward_std": 0.4531472623348236, "rewards/accuracy_reward_step": 0.61328125, "rewards/final_brier_reward_step": 0.7304921746253967, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.8351374864578247, "step": 194 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.4840994033178152, "calib/avg_num_step_conf": 7.8984375, "calib/ece": 0.2280952380952381, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.17063492063492064, "calib/gap": -0.010222280506196335, "calib/mean_conf": 0.8067460317460318, "calib/mu_c": 0.8026490066225165, "calib/mu_w": 0.8128712871287128, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.21781746031746033, "calib/std_conf": 0.10936199802142008, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2443.0, "completions/max_terminated_length": 2443.0, "completions/mean_length": 605.6953125, "completions/mean_terminated_length": 608.0706176757812, "completions/min_length": 0.0, "completions/min_terminated_length": 255.0, "epoch": 0.208, "grad_norm": 0.0071203638799488544, "learning_rate": 1.3888888888888888e-07, "loss": 0.006, "num_tokens": 57299002.0, "reward": 1.7441420555114746, "reward_std": 0.2980097234249115, "rewards/accuracy_reward_step": 0.58984375, "rewards/final_brier_reward_step": 0.6889632940292358, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.7797926068305969, "step": 195 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5241118421052631, "calib/avg_num_step_conf": 7.953125, "calib/ece": 0.21010784313725478, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.99609375, "calib/frac_conf_gt_0.9": 0.2235294117647059, "calib/gap": -0.004050164473684559, "calib/mean_conf": 0.8157745098039215, "calib/mu_c": 0.8142656249999998, "calib/mu_w": 0.8183157894736843, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.1992156862745097, "calib/std_conf": 0.12012087513986715, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 1086.0, "completions/max_terminated_length": 1086.0, "completions/mean_length": 567.0546875, "completions/mean_terminated_length": 569.2784423828125, "completions/min_length": 0.0, "completions/min_terminated_length": 279.0, "epoch": 0.20906666666666668, "grad_norm": 0.006445140112191439, "learning_rate": 1.1111111111111112e-07, "loss": 0.0044, "num_tokens": 57600216.0, "reward": 1.8126583099365234, "reward_std": 0.3135502338409424, "rewards/accuracy_reward_step": 0.625, "rewards/final_brier_reward_step": 0.7116646766662598, "rewards/format_reward_step": 0.99609375, "rewards/stepwise_brier_reward": 0.7967812418937683, "step": 196 }, { "calib/answer_extract_rate": 0.97265625, "calib/auroc": 0.5208441727367326, "calib/avg_num_step_conf": 8.1328125, "calib/ece": 0.3306048387096775, "calib/final_conf_rate": 0.96875, "calib/format_rate": 0.96875, "calib/frac_conf_gt_0.9": 0.22983870967741934, "calib/gap": 0.00911290322580649, "calib/mean_conf": 0.8168951612903225, "calib/mu_c": 0.8214516129032258, "calib/mu_w": 0.8123387096774193, "calib/nonempty_final_conf_rate": 0.96875, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.32375000000000004, "calib/std_conf": 0.123273713190135, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 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": 2775.0, "completions/max_terminated_length": 2775.0, "completions/mean_length": 638.95703125, "completions/mean_terminated_length": 643.9881591796875, "completions/min_length": 0.0, "completions/min_terminated_length": 329.0, "epoch": 0.21013333333333334, "grad_norm": 0.006195712368935347, "learning_rate": 8.333333333333334e-08, "loss": 0.0081, "num_tokens": 57922349.0, "reward": 1.5633504390716553, "reward_std": 0.5351172685623169, "rewards/accuracy_reward_step": 0.48828125, "rewards/final_brier_reward_step": 0.6189706921577454, "rewards/format_reward_step": 0.96875, "rewards/stepwise_brier_reward": 0.7672438025474548, "step": 197 }, { "calib/answer_extract_rate": 0.99609375, "calib/auroc": 0.5231506298579469, "calib/avg_num_step_conf": 8.16015625, "calib/ece": 0.18398235294117649, "calib/final_conf_rate": 0.99609375, "calib/format_rate": 0.9921875, "calib/frac_conf_gt_0.9": 0.18823529411764706, "calib/gap": -0.002887999195926083, "calib/mean_conf": 0.7888019607843137, "calib/mu_c": 0.7877713414634147, "calib/mu_w": 0.7906593406593407, "calib/nonempty_final_conf_rate": 0.99609375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.1648235294117647, "calib/std_conf": 0.13852333630849958, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2016.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 599.5859375, "completions/mean_terminated_length": 601.9373168945312, "completions/min_length": 0.0, "completions/min_terminated_length": 224.0, "epoch": 0.2112, "grad_norm": 0.006844573654234409, "learning_rate": 5.555555555555556e-08, "loss": 0.0389, "num_tokens": 58234731.0, "reward": 1.842313528060913, "reward_std": 0.4178352355957031, "rewards/accuracy_reward_step": 0.640625, "rewards/final_brier_reward_step": 0.723809003829956, "rewards/format_reward_step": 0.9921875, "rewards/stepwise_brier_reward": 0.8173204064369202, "step": 198 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5387755102040817, "calib/avg_num_step_conf": 8.36328125, "calib/ece": 0.2760277777777777, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.984375, "calib/frac_conf_gt_0.9": 0.2222222222222222, "calib/gap": 0.007848214285714361, "calib/mean_conf": 0.8244404761904762, "calib/mu_c": 0.8279285714285715, "calib/mu_w": 0.8200803571428571, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 1.0, "calib/pce": 0.27245634920634915, "calib/std_conf": 0.10445383477772964, "calib/step_conf_rate": 1.0, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00390625, "completions/max_length": 2528.0, "completions/max_terminated_length": 2528.0, "completions/mean_length": 703.8125, "completions/mean_terminated_length": 706.5725708007812, "completions/min_length": 0.0, "completions/min_terminated_length": 241.0, "epoch": 0.21226666666666666, "grad_norm": 0.005577475298196077, "learning_rate": 2.777777777777778e-08, "loss": 0.0211, "num_tokens": 58572611.0, "reward": 1.6749939918518066, "reward_std": 0.5715276598930359, "rewards/accuracy_reward_step": 0.546875, "rewards/final_brier_reward_step": 0.6632249355316162, "rewards/format_reward_step": 0.984375, "rewards/stepwise_brier_reward": 0.7867506146430969, "step": 199 }, { "calib/answer_extract_rate": 0.984375, "calib/auroc": 0.5772438867897627, "calib/avg_num_step_conf": 7.57421875, "calib/ece": 0.13329365079365077, "calib/final_conf_rate": 0.984375, "calib/format_rate": 0.98046875, "calib/frac_conf_gt_0.9": 0.20634920634920634, "calib/gap": 0.03450844799315589, "calib/mean_conf": 0.7991666666666666, "calib/mu_c": 0.8105325443786981, "calib/mu_w": 0.7760240963855423, "calib/nonempty_final_conf_rate": 0.984375, "calib/nonempty_reasoning_rate": 1.0, "calib/nonempty_step_conf_rate": 0.99609375, "calib/pce": 0.13091269841269837, "calib/std_conf": 0.11266830444005459, "calib/step_conf_rate": 0.99609375, "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.015625, "completions/max_length": 2436.0, "completions/max_terminated_length": 2436.0, "completions/mean_length": 643.203125, "completions/mean_terminated_length": 653.4127197265625, "completions/min_length": 0.0, "completions/min_terminated_length": 183.0, "epoch": 0.21333333333333335, "grad_norm": 0.006515799555927515, "learning_rate": 0.0, "loss": -0.0229, "num_tokens": 58898823.0, "reward": 1.8704109191894531, "reward_std": 0.34547263383865356, "rewards/accuracy_reward_step": 0.66015625, "rewards/final_brier_reward_step": 0.7496362924575806, "rewards/format_reward_step": 0.98046875, "rewards/stepwise_brier_reward": 0.8101325035095215, "step": 200 }, { "epoch": 0.21333333333333335, "step": 200, "total_flos": 0.0, "train_loss": 0.025656483445782214, "train_runtime": 13014.1228, "train_samples_per_second": 3.934, "train_steps_per_second": 0.015 } ], "logging_steps": 1, "max_steps": 200, "num_input_tokens_seen": 58898823, "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 }