deepscaleR / gdpo_lambda0.5 /logs /step_stats.jsonl
brikdavies's picture
Add gdpo_lambda0.5/logs
6cd7fb6 verified
Raw
History Blame Contribute Delete
161 kB
{"step": 0, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.8160546875, "mean_total_reward": 0.19197265625, "std_total_reward": 0.5677937558176454, "total_tokens": 83564, "mean_completion_length": 1671.28, "median_completion_length": 1812, "min_completion_length": 992, "max_completion_length": 2048, "frac_has_think": 0.7, "frac_has_boxed": 0.7, "mean_reward_std_per_prompt": 0.23142144295716582, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": null, "avg_step_time_sec": null, "gpu_allocated_gb": null, "gpu_reserved_gb": null, "gpu_peak_gb": null, "elapsed_sec": null, "eta_sec": null, "timestamp": "2026-02-15T12:11:09.893919"}
{"step": 1, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.22, "mean_correctness_reward": 0.22, "mean_length_penalty": -0.94798828125, "mean_total_reward": -0.253994140625, "std_total_reward": 0.4489863735331722, "total_tokens": 97074, "mean_completion_length": 1941.48, "median_completion_length": 2048, "min_completion_length": 1079, "max_completion_length": 2048, "frac_has_think": 0.46, "frac_has_boxed": 0.46, "mean_reward_std_per_prompt": 0.28630213533598464, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 27.942410945892334, "avg_step_time_sec": 27.942410945892334, "gpu_allocated_gb": 32.365621248, "gpu_reserved_gb": 79.92246272, "gpu_peak_gb": 73.906566656, "elapsed_sec": 27.94724130630493, "eta_sec": 5560.5397782325745, "timestamp": "2026-02-15T12:11:39.427386"}
{"step": 2, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.83728515625, "mean_total_reward": 0.201357421875, "std_total_reward": 0.554419128743151, "total_tokens": 85738, "mean_completion_length": 1714.76, "median_completion_length": 1930, "min_completion_length": 753, "max_completion_length": 2048, "frac_has_think": 0.7, "frac_has_boxed": 0.7, "mean_reward_std_per_prompt": 0.4344257708738014, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 29.51460862159729, "avg_step_time_sec": 28.728509783744812, "gpu_allocated_gb": 32.365654016, "gpu_reserved_gb": 79.95392, "gpu_peak_gb": 74.127090176, "elapsed_sec": 57.46796274185181, "eta_sec": 5688.244937181473, "timestamp": "2026-02-15T12:12:07.544043"}
{"step": 3, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.735615234375, "mean_total_reward": 0.3321923828125, "std_total_reward": 0.5507835217368169, "total_tokens": 75327, "mean_completion_length": 1506.54, "median_completion_length": 1562, "min_completion_length": 638, "max_completion_length": 2048, "frac_has_think": 0.74, "frac_has_boxed": 0.74, "mean_reward_std_per_prompt": 0.34958831303817217, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 28.05452299118042, "avg_step_time_sec": 28.503847519556682, "gpu_allocated_gb": 32.365626368, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.127090176, "elapsed_sec": 85.52868008613586, "eta_sec": 5615.257961352667, "timestamp": "2026-02-15T12:12:35.327365"}
{"step": 4, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.24, "mean_correctness_reward": 0.24, "mean_length_penalty": -0.93474609375, "mean_total_reward": -0.227373046875, "std_total_reward": 0.48614390473452884, "total_tokens": 95718, "mean_completion_length": 1914.36, "median_completion_length": 2048, "min_completion_length": 1072, "max_completion_length": 2048, "frac_has_think": 0.24, "frac_has_boxed": 0.24, "mean_reward_std_per_prompt": 0.2018068088814784, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 27.7833092212677, "avg_step_time_sec": 28.323712944984436, "gpu_allocated_gb": 32.365638656, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.127090176, "elapsed_sec": 113.31832933425903, "eta_sec": 5551.4477372169495, "timestamp": "2026-02-15T12:13:04.300153"}
{"step": 5, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.81576171875, "mean_total_reward": 0.172119140625, "std_total_reward": 0.576917940899411, "total_tokens": 83534, "mean_completion_length": 1670.68, "median_completion_length": 1947, "min_completion_length": 764, "max_completion_length": 2048, "frac_has_think": 0.64, "frac_has_boxed": 0.64, "mean_reward_std_per_prompt": 0.28905171090489323, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 29.51466393470764, "avg_step_time_sec": 28.561903142929076, "gpu_allocated_gb": 32.365667328, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.156304384, "elapsed_sec": 142.8398561477661, "eta_sec": 5569.57111287117, "timestamp": "2026-02-15T12:13:32.789486"}
{"step": 6, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.782509765625, "mean_total_reward": 0.3087451171875, "std_total_reward": 0.5342621084839483, "total_tokens": 80129, "mean_completion_length": 1602.58, "median_completion_length": 1741, "min_completion_length": 862, "max_completion_length": 2048, "frac_has_think": 0.72, "frac_has_boxed": 0.72, "mean_reward_std_per_prompt": 0.3923693213718517, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.904213666915894, "avg_step_time_sec": 28.452288230260212, "gpu_allocated_gb": 32.365632512, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.156304384, "elapsed_sec": 170.75053071975708, "eta_sec": 5519.743916670481, "timestamp": "2026-02-15T12:14:00.422556"}
{"step": 7, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.46, "mean_correctness_reward": 0.46, "mean_length_penalty": -0.87857421875, "mean_total_reward": 0.020712890625, "std_total_reward": 0.5663336921467729, "total_tokens": 89966, "mean_completion_length": 1799.32, "median_completion_length": 2048, "min_completion_length": 766, "max_completion_length": 2048, "frac_has_think": 0.46, "frac_has_boxed": 0.46, "mean_reward_std_per_prompt": 0.1914567800054596, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 27.619283437728882, "avg_step_time_sec": 28.33328754561288, "gpu_allocated_gb": 32.36562432, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.156304384, "elapsed_sec": 198.3762743473053, "eta_sec": 5468.324496303286, "timestamp": "2026-02-15T12:14:29.330737"}
{"step": 8, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.64, "mean_correctness_reward": 0.64, "mean_length_penalty": -0.792470703125, "mean_total_reward": 0.2437646484375, "std_total_reward": 0.5251462464995756, "total_tokens": 81149, "mean_completion_length": 1622.98, "median_completion_length": 1743, "min_completion_length": 873, "max_completion_length": 2048, "frac_has_think": 0.82, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.312602602251628, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.8109450340271, "avg_step_time_sec": 28.392994731664658, "gpu_allocated_gb": 32.365616128, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.156304384, "elapsed_sec": 227.19279098510742, "eta_sec": 5451.454988479614, "timestamp": "2026-02-15T12:14:56.971848"}
{"step": 9, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.74587890625, "mean_total_reward": 0.287060546875, "std_total_reward": 0.5550777620533153, "total_tokens": 76378, "mean_completion_length": 1527.56, "median_completion_length": 1550, "min_completion_length": 650, "max_completion_length": 2048, "frac_has_think": 0.82, "frac_has_boxed": 0.82, "mean_reward_std_per_prompt": 0.30030333532198084, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.189920663833618, "avg_step_time_sec": 28.370430946350098, "gpu_allocated_gb": 32.365641728, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.156304384, "elapsed_sec": 255.39127564430237, "eta_sec": 5418.752310752869, "timestamp": "2026-02-15T12:15:24.938451"}
{"step": 10, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.930771484375, "mean_total_reward": 0.0146142578125, "std_total_reward": 0.5369351080262947, "total_tokens": 95311, "mean_completion_length": 1906.22, "median_completion_length": 2048, "min_completion_length": 1323, "max_completion_length": 2048, "frac_has_think": 0.6, "frac_has_boxed": 0.6, "mean_reward_std_per_prompt": 0.498886886244055, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.877158880233765, "avg_step_time_sec": 28.321103739738465, "gpu_allocated_gb": 32.365712384, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 283.2755389213562, "eta_sec": 5381.009710550308, "timestamp": "2026-02-15T12:15:54.290653"}
{"step": 11, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.85080078125, "mean_total_reward": 0.074599609375, "std_total_reward": 0.5696424653734841, "total_tokens": 87122, "mean_completion_length": 1742.44, "median_completion_length": 1907, "min_completion_length": 974, "max_completion_length": 2048, "frac_has_think": 0.7, "frac_has_boxed": 0.7, "mean_reward_std_per_prompt": 0.4631215406199195, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 29.468507766723633, "avg_step_time_sec": 28.425413196737114, "gpu_allocated_gb": 32.36563456, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 312.75112771987915, "eta_sec": 5372.403094183314, "timestamp": "2026-02-15T12:16:23.036372"}
{"step": 12, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.46, "mean_correctness_reward": 0.46, "mean_length_penalty": -0.840888671875, "mean_total_reward": 0.0395556640625, "std_total_reward": 0.5614515419539254, "total_tokens": 86107, "mean_completion_length": 1722.14, "median_completion_length": 1854, "min_completion_length": 936, "max_completion_length": 2048, "frac_has_think": 0.7, "frac_has_boxed": 0.7, "mean_reward_std_per_prompt": 0.3338630169124101, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.277159690856934, "avg_step_time_sec": 28.41305873791377, "gpu_allocated_gb": 32.365643776, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 341.0345160961151, "eta_sec": 5341.655042727789, "timestamp": "2026-02-15T12:16:51.264013"}
{"step": 13, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.83369140625, "mean_total_reward": 0.163154296875, "std_total_reward": 0.5666046069700179, "total_tokens": 85370, "mean_completion_length": 1707.4, "median_completion_length": 1862, "min_completion_length": 939, "max_completion_length": 2048, "frac_has_think": 0.6, "frac_has_boxed": 0.6, "mean_reward_std_per_prompt": 0.4187255449570461, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 28.213676929473877, "avg_step_time_sec": 28.397721675726082, "gpu_allocated_gb": 32.365641728, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 369.25430607795715, "eta_sec": 5310.373953360778, "timestamp": "2026-02-15T12:17:19.844417"}
{"step": 14, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.764970703125, "mean_total_reward": 0.0975146484375, "std_total_reward": 0.5405416547664068, "total_tokens": 78333, "mean_completion_length": 1566.66, "median_completion_length": 1539, "min_completion_length": 923, "max_completion_length": 2048, "frac_has_think": 0.78, "frac_has_boxed": 0.78, "mean_reward_std_per_prompt": 0.2777906211928455, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 28.612775564193726, "avg_step_time_sec": 28.413082667759486, "gpu_allocated_gb": 32.365654016, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 397.8740568161011, "eta_sec": 5284.833376203264, "timestamp": "2026-02-15T12:17:47.409623"}
{"step": 15, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.74681640625, "mean_total_reward": 0.346591796875, "std_total_reward": 0.5332053020502717, "total_tokens": 76474, "mean_completion_length": 1529.48, "median_completion_length": 1522, "min_completion_length": 648, "max_completion_length": 2048, "frac_has_think": 0.76, "frac_has_boxed": 0.76, "mean_reward_std_per_prompt": 0.25102333039749486, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.140949249267578, "avg_step_time_sec": 28.394940439860026, "gpu_allocated_gb": 32.36562944, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 426.02150797843933, "eta_sec": 5253.0639813741045, "timestamp": "2026-02-15T12:18:15.428224"}
{"step": 16, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.44, "mean_correctness_reward": 0.44, "mean_length_penalty": -0.86373046875, "mean_total_reward": 0.008134765625, "std_total_reward": 0.5771310467418754, "total_tokens": 88446, "mean_completion_length": 1768.92, "median_completion_length": 2048, "min_completion_length": 865, "max_completion_length": 2048, "frac_has_think": 0.54, "frac_has_boxed": 0.54, "mean_reward_std_per_prompt": 0.22064636467881635, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 27.371922254562378, "avg_step_time_sec": 28.331001803278923, "gpu_allocated_gb": 32.365674496, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 453.4010043144226, "eta_sec": 5212.904331803322, "timestamp": "2026-02-15T12:18:43.831517"}
{"step": 17, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.852578125, "mean_total_reward": 0.1537109375, "std_total_reward": 0.554443301323065, "total_tokens": 87304, "mean_completion_length": 1746.08, "median_completion_length": 1880, "min_completion_length": 1028, "max_completion_length": 2048, "frac_has_think": 0.84, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.28205092208612625, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.395580291748047, "avg_step_time_sec": 28.334800537894754, "gpu_allocated_gb": 32.365631488, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 481.80290174484253, "eta_sec": 5185.26849843474, "timestamp": "2026-02-15T12:19:12.643844"}
{"step": 18, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.64, "mean_correctness_reward": 0.64, "mean_length_penalty": -0.823369140625, "mean_total_reward": 0.2283154296875, "std_total_reward": 0.5507823466621465, "total_tokens": 84313, "mean_completion_length": 1686.26, "median_completion_length": 1812, "min_completion_length": 740, "max_completion_length": 2048, "frac_has_think": 0.66, "frac_has_boxed": 0.66, "mean_reward_std_per_prompt": 0.2973654911811978, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.825542449951172, "avg_step_time_sec": 28.362063977453445, "gpu_allocated_gb": 32.36563456, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 510.63551139831543, "eta_sec": 5161.895643896527, "timestamp": "2026-02-15T12:19:40.761894"}
{"step": 19, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.87302734375, "mean_total_reward": 0.123486328125, "std_total_reward": 0.5552752709132038, "total_tokens": 89398, "mean_completion_length": 1787.96, "median_completion_length": 1979, "min_completion_length": 775, "max_completion_length": 2048, "frac_has_think": 0.66, "frac_has_boxed": 0.68, "mean_reward_std_per_prompt": 0.35910705052120534, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.15167784690857, "avg_step_time_sec": 28.350991023214238, "gpu_allocated_gb": 32.365650944, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 538.7932739257812, "eta_sec": 5131.529375201777, "timestamp": "2026-02-15T12:20:09.794515"}
{"step": 20, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.834736328125, "mean_total_reward": 0.2826318359375, "std_total_reward": 0.5087758398640582, "total_tokens": 85477, "mean_completion_length": 1709.54, "median_completion_length": 1777, "min_completion_length": 1061, "max_completion_length": 2048, "frac_has_think": 0.76, "frac_has_boxed": 0.76, "mean_reward_std_per_prompt": 0.2876789804188337, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.861501693725586, "avg_step_time_sec": 28.37651655673981, "gpu_allocated_gb": 32.365620224, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 567.6614685058594, "eta_sec": 5107.772980213165, "timestamp": "2026-02-15T12:20:37.860972"}
{"step": 21, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.753798828125, "mean_total_reward": 0.3231005859375, "std_total_reward": 0.5428872379949943, "total_tokens": 77189, "mean_completion_length": 1543.78, "median_completion_length": 1512, "min_completion_length": 758, "max_completion_length": 2048, "frac_has_think": 0.78, "frac_has_boxed": 0.78, "mean_reward_std_per_prompt": 0.32915775800096125, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 28.704607009887695, "avg_step_time_sec": 28.39213991165161, "gpu_allocated_gb": 32.365652992, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 596.3738114833832, "eta_sec": 5082.193044185638, "timestamp": "2026-02-15T12:21:06.066998"}
{"step": 22, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.812509765625, "mean_total_reward": 0.1537451171875, "std_total_reward": 0.5563414055163614, "total_tokens": 83201, "mean_completion_length": 1664.02, "median_completion_length": 1759, "min_completion_length": 824, "max_completion_length": 2048, "frac_has_think": 0.7, "frac_has_boxed": 0.7, "mean_reward_std_per_prompt": 0.37189168371255316, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.24735999107361, "avg_step_time_sec": 28.385559006170794, "gpu_allocated_gb": 32.365769728, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.392025088, "elapsed_sec": 624.6273694038391, "eta_sec": 5052.629503098401, "timestamp": "2026-02-15T12:21:34.890338"}
{"step": 23, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.776640625, "mean_total_reward": 0.1716796875, "std_total_reward": 0.5997526056154362, "total_tokens": 79528, "mean_completion_length": 1590.56, "median_completion_length": 1895, "min_completion_length": 672, "max_completion_length": 2048, "frac_has_think": 0.72, "frac_has_boxed": 0.72, "mean_reward_std_per_prompt": 0.27254382484904865, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 29.161810636520386, "avg_step_time_sec": 28.41930907705556, "gpu_allocated_gb": 32.365832192, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 653.7962987422943, "eta_sec": 5030.217706638834, "timestamp": "2026-02-15T12:22:03.936677"}
{"step": 24, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.704580078125, "mean_total_reward": 0.4077099609375, "std_total_reward": 0.514457702593181, "total_tokens": 72149, "mean_completion_length": 1442.98, "median_completion_length": 1388, "min_completion_length": 633, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.25795185383673036, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 28.003376722335815, "avg_step_time_sec": 28.40197856227557, "gpu_allocated_gb": 32.365621248, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 681.8060436248779, "eta_sec": 4998.748226960501, "timestamp": "2026-02-15T12:22:30.792651"}
{"step": 25, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.616328125, "mean_total_reward": 0.2518359375, "std_total_reward": 0.48304603691343023, "total_tokens": 63112, "mean_completion_length": 1262.24, "median_completion_length": 1206, "min_completion_length": 687, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.14688984186684304, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 26.932655572891235, "avg_step_time_sec": 28.343205642700195, "gpu_allocated_gb": 32.365627392, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 708.7456998825073, "eta_sec": 4960.060987472534, "timestamp": "2026-02-15T12:22:57.418502"}
{"step": 26, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.751669921875, "mean_total_reward": 0.2241650390625, "std_total_reward": 0.5966489067866493, "total_tokens": 76971, "mean_completion_length": 1539.42, "median_completion_length": 1711, "min_completion_length": 202, "max_completion_length": 2048, "frac_has_think": 0.72, "frac_has_boxed": 0.72, "mean_reward_std_per_prompt": 0.15194991416608875, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 26.705711364746094, "avg_step_time_sec": 28.280225093548115, "gpu_allocated_gb": 32.365635584, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 735.4572520256042, "eta_sec": 4920.759166277372, "timestamp": "2026-02-15T12:23:24.856397"}
{"step": 27, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.589580078125, "mean_total_reward": 0.5052099609375, "std_total_reward": 0.40658289140587495, "total_tokens": 60373, "mean_completion_length": 1207.46, "median_completion_length": 1158, "min_completion_length": 638, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.04847309600130973, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 27.24646830558777, "avg_step_time_sec": 28.241937805105138, "gpu_allocated_gb": 32.3656192, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 762.7096374034882, "eta_sec": 4885.855240283189, "timestamp": "2026-02-15T12:23:51.123650"}
{"step": 28, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.7839453125, "mean_total_reward": 0.30802734375, "std_total_reward": 0.5396682993836864, "total_tokens": 80276, "mean_completion_length": 1605.52, "median_completion_length": 1925, "min_completion_length": 545, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.3150212029948824, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.32823920249939, "avg_step_time_sec": 28.173591426440648, "gpu_allocated_gb": 32.365636608, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 789.04350233078, "eta_sec": 4845.8577253477915, "timestamp": "2026-02-15T12:24:18.823998"}
{"step": 29, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.88, "mean_correctness_reward": 0.88, "mean_length_penalty": -0.593623046875, "mean_total_reward": 0.5831884765625, "std_total_reward": 0.40526415949117217, "total_tokens": 60787, "mean_completion_length": 1215.74, "median_completion_length": 1089, "min_completion_length": 676, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.21237646779614677, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 28.01240301132202, "avg_step_time_sec": 28.168033205229662, "gpu_allocated_gb": 32.365682688, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 817.0620815753937, "eta_sec": 4816.733678094272, "timestamp": "2026-02-15T12:24:44.986165"}
{"step": 30, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.42, "mean_correctness_reward": 0.42, "mean_length_penalty": -0.71478515625, "mean_total_reward": 0.062607421875, "std_total_reward": 0.5052808578294342, "total_tokens": 73194, "mean_completion_length": 1463.88, "median_completion_length": 1490, "min_completion_length": 510, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.32102734351247886, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 25.737329721450806, "avg_step_time_sec": 28.087009755770364, "gpu_allocated_gb": 32.365612032, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 842.8054485321045, "eta_sec": 4774.791658480962, "timestamp": "2026-02-15T12:25:12.449646"}
{"step": 31, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.581708984375, "mean_total_reward": 0.5291455078125, "std_total_reward": 0.47628871092075115, "total_tokens": 59567, "mean_completion_length": 1191.34, "median_completion_length": 1099, "min_completion_length": 552, "max_completion_length": 2048, "frac_has_think": 0.88, "frac_has_boxed": 0.88, "mean_reward_std_per_prompt": 0.3377780869741762, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.700984477996826, "avg_step_time_sec": 28.07455732745509, "gpu_allocated_gb": 32.365669376, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 870.5120720863342, "eta_sec": 4744.60018833991, "timestamp": "2026-02-15T12:25:38.438828"}
{"step": 32, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.674833984375, "mean_total_reward": 0.2425830078125, "std_total_reward": 0.5555428712003455, "total_tokens": 69103, "mean_completion_length": 1382.06, "median_completion_length": 1300, "min_completion_length": 706, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.82, "mean_reward_std_per_prompt": 0.19762566695536155, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.744999885559082, "avg_step_time_sec": 28.00175865739584, "gpu_allocated_gb": 32.365612032, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 896.2633438110352, "eta_sec": 4704.295454442501, "timestamp": "2026-02-15T12:26:05.596390"}
{"step": 33, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.643154296875, "mean_total_reward": 0.4384228515625, "std_total_reward": 0.5165385904334642, "total_tokens": 65859, "mean_completion_length": 1317.18, "median_completion_length": 1232, "min_completion_length": 647, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.14959208845545302, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 27.393394470214844, "avg_step_time_sec": 27.983323378996417, "gpu_allocated_gb": 32.365663232, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 923.6625952720642, "eta_sec": 4673.215004292401, "timestamp": "2026-02-15T12:26:32.153857"}
{"step": 34, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.656806640625, "mean_total_reward": 0.4715966796875, "std_total_reward": 0.48659109486056445, "total_tokens": 67257, "mean_completion_length": 1345.14, "median_completion_length": 1175, "min_completion_length": 632, "max_completion_length": 2048, "frac_has_think": 0.88, "frac_has_boxed": 0.88, "mean_reward_std_per_prompt": 0.33961621578254697, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.753193378448486, "avg_step_time_sec": 27.947143084862653, "gpu_allocated_gb": 32.365696, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 950.421847820282, "eta_sec": 4639.2257520872, "timestamp": "2026-02-15T12:26:58.982477"}
{"step": 35, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.88, "mean_correctness_reward": 0.88, "mean_length_penalty": -0.5455078125, "mean_total_reward": 0.60724609375, "std_total_reward": 0.4156646768088435, "total_tokens": 55860, "mean_completion_length": 1117.2, "median_completion_length": 996, "min_completion_length": 532, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.1364034071014737, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.509282112121582, "avg_step_time_sec": 27.906061342784337, "gpu_allocated_gb": 32.365632512, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 976.937413930893, "eta_sec": 4604.500121559416, "timestamp": "2026-02-15T12:27:25.092959"}
{"step": 36, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.598603515625, "mean_total_reward": 0.2806982421875, "std_total_reward": 0.5626963777224181, "total_tokens": 61297, "mean_completion_length": 1225.94, "median_completion_length": 968, "min_completion_length": 605, "max_completion_length": 2048, "frac_has_think": 0.78, "frac_has_boxed": 0.78, "mean_reward_std_per_prompt": 0.10166450899971313, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 26.165114641189575, "avg_step_time_sec": 27.857701712184483, "gpu_allocated_gb": 32.365651968, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1003.108481168747, "eta_sec": 4568.663080798256, "timestamp": "2026-02-15T12:27:51.283700"}
{"step": 37, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.88, "mean_correctness_reward": 0.88, "mean_length_penalty": -0.628759765625, "mean_total_reward": 0.5656201171875, "std_total_reward": 0.3728144928782199, "total_tokens": 64385, "mean_completion_length": 1287.7, "median_completion_length": 1083, "min_completion_length": 667, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.2159631555325987, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.585747480392456, "avg_step_time_sec": 27.8233245707847, "gpu_allocated_gb": 32.36573184, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1029.700519323349, "eta_sec": 4535.201905037906, "timestamp": "2026-02-15T12:28:18.103491"}
{"step": 38, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.552314453125, "mean_total_reward": 0.3438427734375, "std_total_reward": 0.5874491160048196, "total_tokens": 56557, "mean_completion_length": 1131.14, "median_completion_length": 1087, "min_completion_length": 473, "max_completion_length": 2048, "frac_has_think": 0.84, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.09502195480336867, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 26.907450199127197, "avg_step_time_sec": 27.799222613635816, "gpu_allocated_gb": 32.365658112, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1056.614189863205, "eta_sec": 4503.474063409002, "timestamp": "2026-02-15T12:28:44.271101"}
{"step": 39, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.425244140625, "mean_total_reward": 0.7873779296875, "std_total_reward": 0.05328093251779782, "total_tokens": 43545, "mean_completion_length": 870.9, "median_completion_length": 804, "min_completion_length": 654, "max_completion_length": 1812, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.03173829967106662, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 25.47921085357666, "avg_step_time_sec": 27.739735132608658, "gpu_allocated_gb": 32.365620224, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1082.099717617035, "eta_sec": 4466.097356349994, "timestamp": "2026-02-15T12:29:06.942448"}
{"step": 40, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.59283203125, "mean_total_reward": 0.243583984375, "std_total_reward": 0.5453211677977289, "total_tokens": 60706, "mean_completion_length": 1214.12, "median_completion_length": 1126, "min_completion_length": 598, "max_completion_length": 2048, "frac_has_think": 0.9, "frac_has_boxed": 0.9, "mean_reward_std_per_prompt": 0.14097880200699245, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 21.68672800064087, "avg_step_time_sec": 27.588409954309462, "gpu_allocated_gb": 32.365290496, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1103.7919526100159, "eta_sec": 4414.145592689514, "timestamp": "2026-02-15T12:29:31.946637"}
{"step": 41, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.57537109375, "mean_total_reward": 0.512314453125, "std_total_reward": 0.5067870250701466, "total_tokens": 58918, "mean_completion_length": 1178.36, "median_completion_length": 1021, "min_completion_length": 471, "max_completion_length": 2048, "frac_has_think": 0.84, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.1762506524210812, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.460530519485474, "avg_step_time_sec": 27.56090069980156, "gpu_allocated_gb": 32.365696, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1130.2588140964508, "eta_sec": 4382.183211268448, "timestamp": "2026-02-15T12:29:58.761556"}
{"step": 42, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.379609375, "mean_total_reward": 0.7901953125, "std_total_reward": 0.15371019111655668, "total_tokens": 38872, "mean_completion_length": 777.44, "median_completion_length": 702, "min_completion_length": 481, "max_completion_length": 1500, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09039806995065734, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.520427703857422, "avg_step_time_sec": 27.53612753323146, "gpu_allocated_gb": 32.365651968, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1156.7850182056427, "eta_sec": 4350.708150250571, "timestamp": "2026-02-15T12:30:19.361264"}
{"step": 43, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.510146484375, "mean_total_reward": 0.7249267578125, "std_total_reward": 0.1977314998036353, "total_tokens": 52239, "mean_completion_length": 1044.78, "median_completion_length": 954, "min_completion_length": 441, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.10070149463198812, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 18.15632724761963, "avg_step_time_sec": 27.317992642868397, "gpu_allocated_gb": 32.364838912, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1174.9482338428497, "eta_sec": 4288.924844930339, "timestamp": "2026-02-15T12:30:42.185375"}
{"step": 44, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.493173828125, "mean_total_reward": 0.5334130859375, "std_total_reward": 0.48524126671052087, "total_tokens": 50501, "mean_completion_length": 1010.02, "median_completion_length": 872, "min_completion_length": 430, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.09104522514878091, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.205586671829224, "avg_step_time_sec": 27.26998341625387, "gpu_allocated_gb": 32.365645824, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1200.1598896980286, "eta_sec": 4254.117412935603, "timestamp": "2026-02-15T12:31:07.873642"}
{"step": 45, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.5305859375, "mean_total_reward": 0.53470703125, "std_total_reward": 0.5192844839900429, "total_tokens": 54332, "mean_completion_length": 1086.64, "median_completion_length": 934, "min_completion_length": 463, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.02061220324801782, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.677153825759888, "avg_step_time_sec": 27.23458720313178, "gpu_allocated_gb": 32.365638656, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1225.844125509262, "eta_sec": 4221.361016485426, "timestamp": "2026-02-15T12:31:33.459582"}
{"step": 46, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.92, "mean_correctness_reward": 0.92, "mean_length_penalty": -0.403623046875, "mean_total_reward": 0.7181884765625, "std_total_reward": 0.3364869959799794, "total_tokens": 41331, "mean_completion_length": 826.62, "median_completion_length": 684, "min_completion_length": 500, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.1264309053801482, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 25.657837629318237, "avg_step_time_sec": 27.20031003848366, "gpu_allocated_gb": 32.3656704, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1251.5083870887756, "eta_sec": 4188.847745926484, "timestamp": "2026-02-15T12:31:57.864798"}
{"step": 47, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.68, "mean_correctness_reward": 0.68, "mean_length_penalty": -0.496640625, "mean_total_reward": 0.4316796875, "std_total_reward": 0.5295015742681081, "total_tokens": 50856, "mean_completion_length": 1017.12, "median_completion_length": 842, "min_completion_length": 458, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.20497865259649584, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 24.34249472618103, "avg_step_time_sec": 27.139505457370838, "gpu_allocated_gb": 32.36563456, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1275.8570311069489, "eta_sec": 4152.344334977738, "timestamp": "2026-02-15T12:32:23.629243"}
{"step": 48, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.4746484375, "mean_total_reward": 0.56267578125, "std_total_reward": 0.533306470174576, "total_tokens": 48604, "mean_completion_length": 972.08, "median_completion_length": 825, "min_completion_length": 389, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.0216080545434913, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.730355739593506, "avg_step_time_sec": 27.110148171583813, "gpu_allocated_gb": 32.3656448, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1301.5947806835175, "eta_sec": 4120.74252208074, "timestamp": "2026-02-15T12:32:48.667825"}
{"step": 49, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.44544921875, "mean_total_reward": 0.377275390625, "std_total_reward": 0.6010219303991913, "total_tokens": 45614, "mean_completion_length": 912.28, "median_completion_length": 724, "min_completion_length": 287, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.012248640115906899, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 25.034446477890015, "avg_step_time_sec": 27.067786912528835, "gpu_allocated_gb": 32.365645824, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1326.6349670886993, "eta_sec": 4087.235823791854, "timestamp": "2026-02-15T12:33:13.554282"}
{"step": 50, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.3744921875, "mean_total_reward": 0.81275390625, "std_total_reward": 0.07336899764352832, "total_tokens": 38348, "mean_completion_length": 766.96, "median_completion_length": 724, "min_completion_length": 382, "max_completion_length": 1970, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.03301449111267503, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 24.88783597946167, "avg_step_time_sec": 27.02418789386749, "gpu_allocated_gb": 32.365627392, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1351.5295116901398, "eta_sec": 4053.628184080124, "timestamp": "2026-02-15T12:33:38.078392"}
{"step": 51, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.3418359375, "mean_total_reward": 0.62908203125, "std_total_reward": 0.40166776009301325, "total_tokens": 35004, "mean_completion_length": 700.08, "median_completion_length": 640, "min_completion_length": 454, "max_completion_length": 1427, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.020766156464338124, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.435893774032593, "avg_step_time_sec": 26.953829185635435, "gpu_allocated_gb": 32.365513728, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1375.5735833644867, "eta_sec": 4016.1205486596796, "timestamp": "2026-02-15T12:33:57.355066"}
{"step": 52, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.39421875, "mean_total_reward": 0.562890625, "std_total_reward": 0.4933661985189748, "total_tokens": 40368, "mean_completion_length": 807.36, "median_completion_length": 777, "min_completion_length": 397, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.1196534368449182, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 17.071895837783813, "avg_step_time_sec": 26.763792005869057, "gpu_allocated_gb": 32.364752896, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1392.651603937149, "eta_sec": 3961.0412168686203, "timestamp": "2026-02-15T12:34:19.028050"}
{"step": 53, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.86, "mean_correctness_reward": 0.86, "mean_length_penalty": -0.54263671875, "mean_total_reward": 0.588681640625, "std_total_reward": 0.4338515621746726, "total_tokens": 55566, "mean_completion_length": 1111.32, "median_completion_length": 1055, "min_completion_length": 512, "max_completion_length": 2048, "frac_has_think": 0.88, "frac_has_boxed": 0.88, "mean_reward_std_per_prompt": 0.21366893878179966, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 24.335546255111694, "avg_step_time_sec": 26.717976048307598, "gpu_allocated_gb": 32.365645824, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1416.9940357208252, "eta_sec": 3927.542479101217, "timestamp": "2026-02-15T12:34:44.705558"}
{"step": 54, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.520078125, "mean_total_reward": 0.5799609375, "std_total_reward": 0.46637921340253763, "total_tokens": 53256, "mean_completion_length": 1065.12, "median_completion_length": 840, "min_completion_length": 480, "max_completion_length": 2048, "frac_has_think": 0.88, "frac_has_boxed": 0.88, "mean_reward_std_per_prompt": 0.1803223011983568, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.19668984413147, "avg_step_time_sec": 26.708322600082116, "gpu_allocated_gb": 32.365657088, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1443.1970808506012, "eta_sec": 3899.415099611989, "timestamp": "2026-02-15T12:35:10.727504"}
{"step": 55, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.453369140625, "mean_total_reward": 0.3733154296875, "std_total_reward": 0.5738472044014563, "total_tokens": 46425, "mean_completion_length": 928.5, "median_completion_length": 586, "min_completion_length": 346, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.02512269222464303, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 25.49479913711548, "avg_step_time_sec": 26.686258537119084, "gpu_allocated_gb": 32.365641728, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1468.6990599632263, "eta_sec": 3869.507487882267, "timestamp": "2026-02-15T12:35:35.727139"}
{"step": 56, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.513037109375, "mean_total_reward": 0.5434814453125, "std_total_reward": 0.5316653776327589, "total_tokens": 52535, "mean_completion_length": 1050.7, "median_completion_length": 642, "min_completion_length": 299, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.017146370756752096, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 24.994328260421753, "avg_step_time_sec": 26.656045496463776, "gpu_allocated_gb": 32.365648896, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1493.7001140117645, "eta_sec": 3838.4705514907837, "timestamp": "2026-02-15T12:36:01.217410"}
{"step": 57, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.585263671875, "mean_total_reward": 0.2873681640625, "std_total_reward": 0.610166672015596, "total_tokens": 59931, "mean_completion_length": 1198.62, "median_completion_length": 1141, "min_completion_length": 466, "max_completion_length": 2048, "frac_has_think": 0.76, "frac_has_boxed": 0.76, "mean_reward_std_per_prompt": 0.1841280578981834, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.495651245117188, "avg_step_time_sec": 26.6356877025805, "gpu_allocated_gb": 32.365647872, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1519.2023203372955, "eta_sec": 3808.9033414690116, "timestamp": "2026-02-15T12:36:27.925976"}
{"step": 58, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.9, "mean_correctness_reward": 0.9, "mean_length_penalty": -0.330771484375, "mean_total_reward": 0.7346142578125, "std_total_reward": 0.33837655815862117, "total_tokens": 33871, "mean_completion_length": 677.42, "median_completion_length": 697, "min_completion_length": 304, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.17193306136235517, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.154814958572388, "avg_step_time_sec": 26.644638172511396, "gpu_allocated_gb": 32.36573696, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1546.3635885715485, "eta_sec": 3783.538620496618, "timestamp": "2026-02-15T12:36:51.821716"}
{"step": 59, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.38, "mean_correctness_reward": 0.38, "mean_length_penalty": -0.3616796875, "mean_total_reward": 0.19916015625, "std_total_reward": 0.4837587745127005, "total_tokens": 37036, "mean_completion_length": 740.72, "median_completion_length": 784, "min_completion_length": 340, "max_completion_length": 1135, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0815992949259785, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 23.344703197479248, "avg_step_time_sec": 26.588707071239664, "gpu_allocated_gb": 32.365630464, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1569.7144339084625, "eta_sec": 3749.0076970447926, "timestamp": "2026-02-15T12:37:09.924952"}
{"step": 60, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.29693359375, "mean_total_reward": 0.851533203125, "std_total_reward": 0.024502994276529248, "total_tokens": 30406, "mean_completion_length": 608.12, "median_completion_length": 612, "min_completion_length": 399, "max_completion_length": 896, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.017402257515176486, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 14.660086870193481, "avg_step_time_sec": 26.389896734555563, "gpu_allocated_gb": 32.364374016, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1584.380390882492, "eta_sec": 3694.585542837779, "timestamp": "2026-02-15T12:37:22.961660"}
{"step": 61, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.434873046875, "mean_total_reward": 0.6225634765625, "std_total_reward": 0.4013110178308307, "total_tokens": 44531, "mean_completion_length": 890.62, "median_completion_length": 858, "min_completion_length": 326, "max_completion_length": 1995, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.26482610419961833, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 11.903944730758667, "avg_step_time_sec": 26.1524221115425, "gpu_allocated_gb": 32.364007424, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1596.2907721996307, "eta_sec": 3635.1866735044073, "timestamp": "2026-02-15T12:37:42.501963"}
{"step": 62, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.345205078125, "mean_total_reward": 0.6273974609375, "std_total_reward": 0.48753323732681003, "total_tokens": 35349, "mean_completion_length": 706.98, "median_completion_length": 620, "min_completion_length": 219, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.02532089378600415, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 24.01100516319275, "avg_step_time_sec": 26.117883128504598, "gpu_allocated_gb": 32.365586432, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1620.3074553012848, "eta_sec": 3604.2678717336344, "timestamp": "2026-02-15T12:38:06.274241"}
{"step": 63, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.341943359375, "mean_total_reward": 0.7890283203125, "std_total_reward": 0.20254749076765552, "total_tokens": 35015, "mean_completion_length": 700.3, "median_completion_length": 668, "min_completion_length": 324, "max_completion_length": 1117, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.13068854105202496, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 24.011967420578003, "avg_step_time_sec": 26.084455895045448, "gpu_allocated_gb": 32.365652992, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1644.32559466362, "eta_sec": 3573.5704576212265, "timestamp": "2026-02-15T12:38:24.292147"}
{"step": 64, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.371513671875, "mean_total_reward": 0.7542431640625, "std_total_reward": 0.27555161504775755, "total_tokens": 38043, "mean_completion_length": 760.86, "median_completion_length": 703, "min_completion_length": 306, "max_completion_length": 1941, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15571960825110345, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 14.226980924606323, "avg_step_time_sec": 25.899182848632336, "gpu_allocated_gb": 32.36434176, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1658.558688879013, "eta_sec": 3522.2888674139977, "timestamp": "2026-02-15T12:38:44.493092"}
{"step": 65, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.370107421875, "mean_total_reward": 0.5749462890625, "std_total_reward": 0.5559250632401873, "total_tokens": 37899, "mean_completion_length": 757.98, "median_completion_length": 525, "min_completion_length": 294, "max_completion_length": 2048, "frac_has_think": 0.88, "frac_has_boxed": 0.88, "mean_reward_std_per_prompt": 0.10076185194499919, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.573325872421265, "avg_step_time_sec": 25.863400433613705, "gpu_allocated_gb": 32.365512192, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1682.137942790985, "eta_sec": 3491.5590585378504, "timestamp": "2026-02-15T12:39:08.516641"}
{"step": 66, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.327734375, "mean_total_reward": 0.4361328125, "std_total_reward": 0.5103879183503746, "total_tokens": 33560, "mean_completion_length": 671.2, "median_completion_length": 683, "min_completion_length": 130, "max_completion_length": 1330, "frac_has_think": 0.98, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.018165618054366308, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 24.395267248153687, "avg_step_time_sec": 25.841155991409764, "gpu_allocated_gb": 32.36565504, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1706.5395765304565, "eta_sec": 3462.7149028489084, "timestamp": "2026-02-15T12:39:27.668146"}
{"step": 67, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.346669921875, "mean_total_reward": 0.6066650390625, "std_total_reward": 0.4686174974723666, "total_tokens": 35499, "mean_completion_length": 709.98, "median_completion_length": 649, "min_completion_length": 377, "max_completion_length": 1327, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.07628336273759814, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 16.177164554595947, "avg_step_time_sec": 25.696917313248363, "gpu_allocated_gb": 32.364633088, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1722.7230508327484, "eta_sec": 3417.6900026620324, "timestamp": "2026-02-15T12:39:44.045273"}
{"step": 68, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.33935546875, "mean_total_reward": 0.430322265625, "std_total_reward": 0.5788851364754041, "total_tokens": 34750, "mean_completion_length": 695, "median_completion_length": 444, "min_completion_length": 219, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.031350310466943435, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 16.38501524925232, "avg_step_time_sec": 25.55997757701313, "gpu_allocated_gb": 32.364639744, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1739.1142086982727, "eta_sec": 3373.917040165733, "timestamp": "2026-02-15T12:40:05.729416"}
{"step": 69, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.47587890625, "mean_total_reward": -0.037939453125, "std_total_reward": 0.48481704230207373, "total_tokens": 48730, "mean_completion_length": 974.6, "median_completion_length": 617, "min_completion_length": 360, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.01697566745658095, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 24.616228580474854, "avg_step_time_sec": 25.546300055324167, "gpu_allocated_gb": 32.365641728, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1763.7367091178894, "eta_sec": 3346.565307247466, "timestamp": "2026-02-15T12:40:30.992891"}
{"step": 70, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.235302734375, "mean_total_reward": 0.3823486328125, "std_total_reward": 0.5113653188398407, "total_tokens": 24095, "mean_completion_length": 481.9, "median_completion_length": 465, "min_completion_length": 269, "max_completion_length": 1306, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.17535963819310843, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 25.268404960632324, "avg_step_time_sec": 25.5423301254, "gpu_allocated_gb": 32.36564992, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1789.011617898941, "eta_sec": 3320.502916302, "timestamp": "2026-02-15T12:40:49.013547"}
{"step": 71, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.236279296875, "mean_total_reward": 0.6218603515625, "std_total_reward": 0.4727655937138584, "total_tokens": 24195, "mean_completion_length": 483.9, "median_completion_length": 446, "min_completion_length": 178, "max_completion_length": 1143, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.10479104576438544, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 14.804280519485474, "avg_step_time_sec": 25.391089990105428, "gpu_allocated_gb": 32.364584448, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1803.8224890232086, "eta_sec": 3275.4506087236, "timestamp": "2026-02-15T12:41:03.014421"}
{"step": 72, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.3533203125, "mean_total_reward": 0.40333984375, "std_total_reward": 0.5715288518362767, "total_tokens": 36180, "mean_completion_length": 723.6, "median_completion_length": 487, "min_completion_length": 261, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.15693012297711526, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 13.590783596038818, "avg_step_time_sec": 25.227196845743393, "gpu_allocated_gb": 32.36439808, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1817.4194867610931, "eta_sec": 3229.0811962551543, "timestamp": "2026-02-15T12:41:24.062555"}
{"step": 73, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.2463671875, "mean_total_reward": 0.27681640625, "std_total_reward": 0.5053432382117842, "total_tokens": 25228, "mean_completion_length": 504.56, "median_completion_length": 484, "min_completion_length": 288, "max_completion_length": 916, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.016935434673223896, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 24.63707184791565, "avg_step_time_sec": 25.21911294166356, "gpu_allocated_gb": 32.36562944, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1842.0625839233398, "eta_sec": 3202.827343591272, "timestamp": "2026-02-15T12:41:39.935024"}
{"step": 74, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.2098046875, "mean_total_reward": 0.73509765625, "std_total_reward": 0.37011857877686233, "total_tokens": 21484, "mean_completion_length": 429.68, "median_completion_length": 344, "min_completion_length": 153, "max_completion_length": 1005, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1556635719931365, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 11.371514320373535, "avg_step_time_sec": 25.031983230565046, "gpu_allocated_gb": 32.364073984, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1853.4393503665924, "eta_sec": 3154.029887051196, "timestamp": "2026-02-15T12:41:51.614684"}
{"step": 75, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.288974609375, "mean_total_reward": 0.6355126953125, "std_total_reward": 0.5420002669768362, "total_tokens": 29591, "mean_completion_length": 591.82, "median_completion_length": 370, "min_completion_length": 148, "max_completion_length": 2048, "frac_has_think": 0.9, "frac_has_boxed": 0.9, "mean_reward_std_per_prompt": 0.08897327847041676, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 11.934244394302368, "avg_step_time_sec": 24.85734671274821, "gpu_allocated_gb": 32.364169728, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1865.3793902397156, "eta_sec": 3107.1683390935264, "timestamp": "2026-02-15T12:42:11.005897"}
{"step": 76, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.2615625, "mean_total_reward": 0.68921875, "std_total_reward": 0.4432878558177965, "total_tokens": 26784, "mean_completion_length": 535.68, "median_completion_length": 365, "min_completion_length": 186, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.20614872777088672, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.63726544380188, "avg_step_time_sec": 24.84129301184102, "gpu_allocated_gb": 32.365657088, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1889.022320508957, "eta_sec": 3080.3203334682867, "timestamp": "2026-02-15T12:42:34.315895"}
{"step": 77, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.216806640625, "mean_total_reward": 0.4915966796875, "std_total_reward": 0.5250678429365221, "total_tokens": 22201, "mean_completion_length": 444.02, "median_completion_length": 391, "min_completion_length": 219, "max_completion_length": 1322, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.025024125923209992, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 23.20442247390747, "avg_step_time_sec": 24.82003495290682, "gpu_allocated_gb": 32.36562432, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1912.232225894928, "eta_sec": 3052.864299207539, "timestamp": "2026-02-15T12:42:52.876587"}
{"step": 78, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.216806640625, "mean_total_reward": 0.6115966796875, "std_total_reward": 0.4396480206115183, "total_tokens": 22201, "mean_completion_length": 444.02, "median_completion_length": 473, "min_completion_length": 216, "max_completion_length": 700, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.15387726033739269, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 15.551734447479248, "avg_step_time_sec": 24.701210587452618, "gpu_allocated_gb": 32.364610048, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1927.789615869522, "eta_sec": 3013.547691669219, "timestamp": "2026-02-15T12:43:04.094716"}
{"step": 79, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.199892578125, "mean_total_reward": 0.3000537109375, "std_total_reward": 0.5081415876329756, "total_tokens": 20469, "mean_completion_length": 409.38, "median_completion_length": 431, "min_completion_length": 155, "max_completion_length": 665, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.012218902933653073, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 9.052194118499756, "avg_step_time_sec": 24.503121771389925, "gpu_allocated_gb": 32.363761664, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1936.8473873138428, "eta_sec": 2964.8777343381807, "timestamp": "2026-02-15T12:43:12.759568"}
{"step": 80, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.203896484375, "mean_total_reward": 0.6580517578125, "std_total_reward": 0.4140417846392115, "total_tokens": 20879, "mean_completion_length": 417.58, "median_completion_length": 365, "min_completion_length": 214, "max_completion_length": 818, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12958782014348405, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.526179313659668, "avg_step_time_sec": 24.303409990668296, "gpu_allocated_gb": 32.363715584, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1945.3790309429169, "eta_sec": 2916.4091988801956, "timestamp": "2026-02-15T12:43:22.410123"}
{"step": 81, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.68, "mean_correctness_reward": 0.68, "mean_length_penalty": -0.143828125, "mean_total_reward": 0.6080859375, "std_total_reward": 0.4895518065129275, "total_tokens": 14728, "mean_completion_length": 294.56, "median_completion_length": 258, "min_completion_length": 119, "max_completion_length": 609, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.16574462136254714, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 10.470338821411133, "avg_step_time_sec": 24.132631334257713, "gpu_allocated_gb": 32.364015616, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1955.8550717830658, "eta_sec": 2871.783128776668, "timestamp": "2026-02-15T12:43:31.033462"}
{"step": 82, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.11587890625, "mean_total_reward": 0.522060546875, "std_total_reward": 0.4969020566471289, "total_tokens": 11866, "mean_completion_length": 237.32, "median_completion_length": 225, "min_completion_length": 109, "max_completion_length": 491, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.06770912364770802, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 7.547680616378784, "avg_step_time_sec": 23.930375837698215, "gpu_allocated_gb": 32.3636224, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1963.409287929535, "eta_sec": 2823.7843488483895, "timestamp": "2026-02-15T12:43:37.599893"}
{"step": 83, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.18318359375, "mean_total_reward": 0.648408203125, "std_total_reward": 0.43457515040758615, "total_tokens": 18758, "mean_completion_length": 375.16, "median_completion_length": 205, "min_completion_length": 81, "max_completion_length": 1617, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15190000498650977, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 6.137210130691528, "avg_step_time_sec": 23.71600034725235, "gpu_allocated_gb": 32.363451392, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1969.5518958568573, "eta_sec": 2774.7720406285252, "timestamp": "2026-02-15T12:43:51.811350"}
{"step": 84, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.1771484375, "mean_total_reward": 0.47142578125, "std_total_reward": 0.5114305129037191, "total_tokens": 18140, "mean_completion_length": 362.8, "median_completion_length": 250, "min_completion_length": 144, "max_completion_length": 1556, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.16530225652454006, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 18.4716854095459, "avg_step_time_sec": 23.65356802656537, "gpu_allocated_gb": 32.365058048, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1988.0301945209503, "eta_sec": 2743.813891081583, "timestamp": "2026-02-15T12:44:09.136184"}
{"step": 85, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.08958984375, "mean_total_reward": 0.895205078125, "std_total_reward": 0.23718356364204177, "total_tokens": 9174, "mean_completion_length": 183.48, "median_completion_length": 182, "min_completion_length": 57, "max_completion_length": 332, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09623191902727368, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 16.847084999084473, "avg_step_time_sec": 23.57349175565383, "gpu_allocated_gb": 32.364919808, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2004.8832767009735, "eta_sec": 2710.95155190019, "timestamp": "2026-02-15T12:44:17.890302"}
{"step": 86, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.137568359375, "mean_total_reward": 0.4912158203125, "std_total_reward": 0.5144602862462864, "total_tokens": 14087, "mean_completion_length": 281.74, "median_completion_length": 228, "min_completion_length": 74, "max_completion_length": 630, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1279164195322012, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 4.35907244682312, "avg_step_time_sec": 23.350068275318588, "gpu_allocated_gb": 32.363211776, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2009.2498593330383, "eta_sec": 2661.907783386319, "timestamp": "2026-02-15T12:44:24.474856"}
{"step": 87, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.0889453125, "mean_total_reward": 0.73552734375, "std_total_reward": 0.4212357036983225, "total_tokens": 9108, "mean_completion_length": 182.16, "median_completion_length": 160, "min_completion_length": 41, "max_completion_length": 400, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2687917032915423, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.715717315673828, "avg_step_time_sec": 23.170363091874396, "gpu_allocated_gb": 32.363636224, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2016.971227645874, "eta_sec": 2618.251029381807, "timestamp": "2026-02-15T12:44:30.444666"}
{"step": 88, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.44, "mean_correctness_reward": 0.44, "mean_length_penalty": -0.10400390625, "mean_total_reward": 0.387998046875, "std_total_reward": 0.5084367748553094, "total_tokens": 10650, "mean_completion_length": 213, "median_completion_length": 181, "min_completion_length": 63, "max_completion_length": 512, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.20276428965688495, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 5.185534238815308, "avg_step_time_sec": 22.965990036725998, "gpu_allocated_gb": 32.36334848, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2022.1623060703278, "eta_sec": 2572.1908841133118, "timestamp": "2026-02-15T12:44:36.431579"}
{"step": 89, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.12630859375, "mean_total_reward": 0.596845703125, "std_total_reward": 0.487937001304914, "total_tokens": 12934, "mean_completion_length": 258.68, "median_completion_length": 205, "min_completion_length": 109, "max_completion_length": 1107, "frac_has_think": 0.96, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.3384565489740345, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.40971827507019, "avg_step_time_sec": 22.779964511314134, "gpu_allocated_gb": 32.363521024, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2028.5771253108978, "eta_sec": 2528.5760607558686, "timestamp": "2026-02-15T12:44:46.580922"}
{"step": 90, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.64, "mean_correctness_reward": 0.64, "mean_length_penalty": -0.082060546875, "mean_total_reward": 0.5989697265625, "std_total_reward": 0.497330040702305, "total_tokens": 8403, "mean_completion_length": 168.06, "median_completion_length": 102, "min_completion_length": 43, "max_completion_length": 784, "frac_has_think": 0.98, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08450988329055743, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 12.20114016532898, "avg_step_time_sec": 22.662422018580965, "gpu_allocated_gb": 32.36432128, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2040.7840030193329, "eta_sec": 2492.8664220439064, "timestamp": "2026-02-15T12:44:56.791304"}
{"step": 91, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.066494140625, "mean_total_reward": 0.5467529296875, "std_total_reward": 0.49678781872176564, "total_tokens": 6809, "mean_completion_length": 136.18, "median_completion_length": 123, "min_completion_length": 51, "max_completion_length": 292, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.14502804428259425, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 9.587255477905273, "avg_step_time_sec": 22.518738869782332, "gpu_allocated_gb": 32.364023296, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2050.376510620117, "eta_sec": 2454.5425368062743, "timestamp": "2026-02-15T12:45:03.446839"}
{"step": 92, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.1290625, "mean_total_reward": 0.63546875, "std_total_reward": 0.5253508999125636, "total_tokens": 13216, "mean_completion_length": 264.32, "median_completion_length": 115, "min_completion_length": 20, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.14501825501346968, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 4.410950183868408, "avg_step_time_sec": 22.32191507971805, "gpu_allocated_gb": 32.363170304, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2054.7928335666656, "eta_sec": 2410.7668286095495, "timestamp": "2026-02-15T12:45:19.246502"}
{"step": 93, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.52, "mean_correctness_reward": 0.52, "mean_length_penalty": -0.107578125, "mean_total_reward": 0.4662109375, "std_total_reward": 0.49426331455741185, "total_tokens": 11016, "mean_completion_length": 220.32, "median_completion_length": 118, "min_completion_length": 58, "max_completion_length": 1413, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1678416419481565, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 22.973748683929443, "avg_step_time_sec": 22.328924043204196, "gpu_allocated_gb": 32.365775872, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2077.7732965946198, "eta_sec": 2389.194872622849, "timestamp": "2026-02-15T12:45:37.750794"}
{"step": 94, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.047587890625, "mean_total_reward": 0.6762060546875, "std_total_reward": 0.45923715666814663, "total_tokens": 4873, "mean_completion_length": 97.46, "median_completion_length": 91, "min_completion_length": 36, "max_completion_length": 194, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.37149231680055206, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 16.10826015472412, "avg_step_time_sec": 22.262746767794834, "gpu_allocated_gb": 32.36493312, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2093.8884654045105, "eta_sec": 2359.8511573862525, "timestamp": "2026-02-15T12:45:45.737348"}
{"step": 95, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.047705078125, "mean_total_reward": 0.5761474609375, "std_total_reward": 0.4981052529023768, "total_tokens": 4885, "mean_completion_length": 97.7, "median_completion_length": 70, "min_completion_length": 26, "max_completion_length": 551, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.20526463629822006, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.837432861328125, "avg_step_time_sec": 22.05826977930571, "gpu_allocated_gb": 32.363026432, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2096.73202252388, "eta_sec": 2316.1183268270997, "timestamp": "2026-02-15T12:45:50.739730"}
{"step": 96, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.042470703125, "mean_total_reward": 0.4587646484375, "std_total_reward": 0.4978577479879529, "total_tokens": 4349, "mean_completion_length": 86.98, "median_completion_length": 75, "min_completion_length": 29, "max_completion_length": 224, "frac_has_think": 1.0, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.35344265635983796, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.352843761444092, "avg_step_time_sec": 21.89467159161965, "gpu_allocated_gb": 32.363558912, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2103.091066837311, "eta_sec": 2277.045845528444, "timestamp": "2026-02-15T12:45:55.110688"}
{"step": 97, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.02712890625, "mean_total_reward": 0.386435546875, "std_total_reward": 0.4932566009056601, "total_tokens": 2778, "mean_completion_length": 55.56, "median_completion_length": 51, "min_completion_length": 18, "max_completion_length": 150, "frac_has_think": 1.0, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.0041079096151590815, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 3.1379683017730713, "avg_step_time_sec": 21.70130351646659, "gpu_allocated_gb": 32.363090944, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2106.2354593276978, "eta_sec": 2235.234262196059, "timestamp": "2026-02-15T12:45:57.656520"}
{"step": 98, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.02486328125, "mean_total_reward": 0.747568359375, "std_total_reward": 0.42916969021716644, "total_tokens": 2546, "mean_completion_length": 50.92, "median_completion_length": 53, "min_completion_length": 22, "max_completion_length": 94, "frac_has_think": 1.0, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.08196869795675316, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.416447162628174, "avg_step_time_sec": 21.504519267958038, "gpu_allocated_gb": 32.362970624, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2108.6572666168213, "eta_sec": 2193.4609653317198, "timestamp": "2026-02-15T12:45:59.760214"}
{"step": 99, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.0258203125, "mean_total_reward": 0.78708984375, "std_total_reward": 0.3999801626054788, "total_tokens": 2644, "mean_completion_length": 52.88, "median_completion_length": 47, "min_completion_length": 22, "max_completion_length": 125, "frac_has_think": 1.0, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.25176096381949986, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.0336413383483887, "avg_step_time_sec": 21.307843733315515, "gpu_allocated_gb": 32.362907136, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2110.6972494125366, "eta_sec": 2152.092217064867, "timestamp": "2026-02-15T12:46:01.970237"}
{"step": 100, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.025927734375, "mean_total_reward": 0.4870361328125, "std_total_reward": 0.5008556179821463, "total_tokens": 2655, "mean_completion_length": 53.1, "median_completion_length": 52, "min_completion_length": 15, "max_completion_length": 244, "frac_has_think": 1.0, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.21142177110486007, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.22053861618042, "avg_step_time_sec": 21.116970682144164, "gpu_allocated_gb": 32.362924544, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2112.9262149333954, "eta_sec": 2111.6970682144165, "timestamp": "2026-02-15T12:46:05.524617"}
{"step": 101, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.02400390625, "mean_total_reward": 0.527998046875, "std_total_reward": 0.5001798674356085, "total_tokens": 2458, "mean_completion_length": 49.16, "median_completion_length": 34, "min_completion_length": 10, "max_completion_length": 198, "frac_has_think": 1.0, "frac_has_boxed": 0.78, "mean_reward_std_per_prompt": 0.1820409146174605, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 3.754300594329834, "avg_step_time_sec": 20.94506305751234, "gpu_allocated_gb": 32.363144192, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2117.285096883774, "eta_sec": 2073.5612426937214, "timestamp": "2026-02-15T12:46:09.013313"}
{"step": 102, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.34, "mean_correctness_reward": 0.34, "mean_length_penalty": -0.065361328125, "mean_total_reward": 0.3073193359375, "std_total_reward": 0.4938883051498109, "total_tokens": 6693, "mean_completion_length": 133.86, "median_completion_length": 54, "min_completion_length": 8, "max_completion_length": 1734, "frac_has_think": 1.0, "frac_has_boxed": 0.76, "mean_reward_std_per_prompt": 0.364397832424982, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.806668758392334, "avg_step_time_sec": 20.767235662422927, "gpu_allocated_gb": 32.363070464, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2120.0981714725494, "eta_sec": 2035.1890949174467, "timestamp": "2026-02-15T12:46:21.930365"}
{"step": 103, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.06, "mean_correctness_reward": 0.06, "mean_length_penalty": -0.02517578125, "mean_total_reward": 0.047412109375, "std_total_reward": 0.2351574058124108, "total_tokens": 2578, "mean_completion_length": 51.56, "median_completion_length": 32, "min_completion_length": 13, "max_completion_length": 222, "frac_has_think": 1.0, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.09619973331995778, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 18.926066160202026, "avg_step_time_sec": 20.74936023036253, "gpu_allocated_gb": 32.365295616, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2139.0289154052734, "eta_sec": 2012.6879423451655, "timestamp": "2026-02-15T12:46:30.919723"}
{"step": 104, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.34, "mean_correctness_reward": 0.34, "mean_length_penalty": -0.011259765625, "mean_total_reward": 0.3343701171875, "std_total_reward": 0.47374882320469264, "total_tokens": 1153, "mean_completion_length": 23.06, "median_completion_length": 19, "min_completion_length": 8, "max_completion_length": 65, "frac_has_think": 1.0, "frac_has_boxed": 0.3, "mean_reward_std_per_prompt": 0.14106837798625357, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 3.035540819168091, "avg_step_time_sec": 20.57903504371643, "gpu_allocated_gb": 32.363100672, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2142.07040643692, "eta_sec": 1975.5873641967773, "timestamp": "2026-02-15T12:46:32.867314"}
{"step": 105, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.38, "mean_correctness_reward": 0.38, "mean_length_penalty": -0.009677734375, "mean_total_reward": 0.3751611328125, "std_total_reward": 0.48431307639258653, "total_tokens": 991, "mean_completion_length": 19.82, "median_completion_length": 19, "min_completion_length": 8, "max_completion_length": 67, "frac_has_think": 1.0, "frac_has_boxed": 0.36, "mean_reward_std_per_prompt": 0.25125018284312156, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 1.714695692062378, "avg_step_time_sec": 20.399374668938773, "gpu_allocated_gb": 32.36284928, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2143.790761947632, "eta_sec": 1937.9405935491834, "timestamp": "2026-02-15T12:46:34.592865"}
{"step": 106, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.00951171875, "mean_total_reward": 0.195244140625, "std_total_reward": 0.4011981247260023, "total_tokens": 974, "mean_completion_length": 19.48, "median_completion_length": 12, "min_completion_length": 8, "max_completion_length": 137, "frac_has_think": 1.0, "frac_has_boxed": 0.3, "mean_reward_std_per_prompt": 0.0027943368309064675, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.737776279449463, "avg_step_time_sec": 20.22332185394359, "gpu_allocated_gb": 32.362860032, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2145.5340061187744, "eta_sec": 1900.9922542706977, "timestamp": "2026-02-15T12:46:36.807694"}
{"step": 107, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.013134765625, "mean_total_reward": 0.3934326171875, "std_total_reward": 0.49307980321830713, "total_tokens": 1345, "mean_completion_length": 26.9, "median_completion_length": 21, "min_completion_length": 10, "max_completion_length": 71, "frac_has_think": 1.0, "frac_has_boxed": 0.12, "mean_reward_std_per_prompt": 0.001844816715713212, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 2.3303935527801514, "avg_step_time_sec": 20.056098224960756, "gpu_allocated_gb": 32.36303872, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2147.8701119422913, "eta_sec": 1865.2171349213504, "timestamp": "2026-02-15T12:46:38.746922"}
{"step": 108, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.52, "mean_correctness_reward": 0.52, "mean_length_penalty": -0.007353515625, "mean_total_reward": 0.5163232421875, "std_total_reward": 0.49954980373678437, "total_tokens": 753, "mean_completion_length": 15.06, "median_completion_length": 12, "min_completion_length": 10, "max_completion_length": 43, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.09803432950708936, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.795090913772583, "avg_step_time_sec": 19.887014823931235, "gpu_allocated_gb": 32.362888192, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2149.6729402542114, "eta_sec": 1829.6053638016735, "timestamp": "2026-02-15T12:46:40.340801"}
{"step": 109, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.36, "mean_correctness_reward": 0.36, "mean_length_penalty": -0.008955078125, "mean_total_reward": 0.3555224609375, "std_total_reward": 0.47809801284821496, "total_tokens": 917, "mean_completion_length": 18.34, "median_completion_length": 12, "min_completion_length": 8, "max_completion_length": 59, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.2794886560393329, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.5735878944396973, "avg_step_time_sec": 19.719001732834982, "gpu_allocated_gb": 32.362812928, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2151.255296945572, "eta_sec": 1794.4291576879834, "timestamp": "2026-02-15T12:46:42.017923"}
{"step": 110, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.00716796875, "mean_total_reward": 0.536416015625, "std_total_reward": 0.49736875403490555, "total_tokens": 734, "mean_completion_length": 14.68, "median_completion_length": 10, "min_completion_length": 8, "max_completion_length": 44, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.09122700636433789, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.68729567527771, "avg_step_time_sec": 19.555077132311734, "gpu_allocated_gb": 32.362845184, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2152.949054479599, "eta_sec": 1759.956941908056, "timestamp": "2026-02-15T12:46:43.618688"}
{"step": 111, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.00626953125, "mean_total_reward": 0.196865234375, "std_total_reward": 0.3996155195164666, "total_tokens": 642, "mean_completion_length": 12.84, "median_completion_length": 14, "min_completion_length": 5, "max_completion_length": 34, "frac_has_think": 0.9, "frac_has_boxed": 0.04, "mean_reward_std_per_prompt": 0.0005091310201836338, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.5783250331878662, "avg_step_time_sec": 19.393124410698007, "gpu_allocated_gb": 32.362842112, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2154.534567594528, "eta_sec": 1725.9880725521227, "timestamp": "2026-02-15T12:46:45.126337"}
{"step": 112, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.18, "mean_correctness_reward": 0.18, "mean_length_penalty": -0.007724609375, "mean_total_reward": 0.1761376953125, "std_total_reward": 0.3848430178809936, "total_tokens": 791, "mean_completion_length": 15.82, "median_completion_length": 14, "min_completion_length": 10, "max_completion_length": 28, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.140131847849703, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.492384672164917, "avg_step_time_sec": 19.233296377318247, "gpu_allocated_gb": 32.362827776, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2156.0335278511047, "eta_sec": 1692.5300812040057, "timestamp": "2026-02-15T12:46:46.606300"}
{"step": 113, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.24, "mean_correctness_reward": 0.24, "mean_length_penalty": -0.00515625, "mean_total_reward": 0.237421875, "std_total_reward": 0.42690848893576133, "total_tokens": 528, "mean_completion_length": 10.56, "median_completion_length": 10, "min_completion_length": 8, "max_completion_length": 12, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.2579078419273158, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.9936292171478271, "avg_step_time_sec": 19.080732951122048, "gpu_allocated_gb": 32.362835456, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2158.032294034958, "eta_sec": 1660.023766747618, "timestamp": "2026-02-15T12:46:48.504508"}
{"step": 114, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.18, "mean_correctness_reward": 0.18, "mean_length_penalty": -0.00671875, "mean_total_reward": 0.176640625, "std_total_reward": 0.38421346864797573, "total_tokens": 688, "mean_completion_length": 13.76, "median_completion_length": 14, "min_completion_length": 6, "max_completion_length": 44, "frac_has_think": 0.98, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.06044675892822683, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.388113260269165, "avg_step_time_sec": 18.92553453278123, "gpu_allocated_gb": 32.36278528, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2159.4263796806335, "eta_sec": 1627.595969819186, "timestamp": "2026-02-15T12:46:50.090832"}
{"step": 115, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.004970703125, "mean_total_reward": 0.3975146484375, "std_total_reward": 0.49010384279159996, "total_tokens": 509, "mean_completion_length": 10.18, "median_completion_length": 10, "min_completion_length": 8, "max_completion_length": 29, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.00038492327864238007, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.5692903995513916, "avg_step_time_sec": 18.774610670753148, "gpu_allocated_gb": 32.362835968, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2161.0022497177124, "eta_sec": 1595.8419070140176, "timestamp": "2026-02-15T12:46:51.572126"}
{"step": 116, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.0055078125, "mean_total_reward": 0.39724609375, "std_total_reward": 0.49017509348834337, "total_tokens": 564, "mean_completion_length": 11.28, "median_completion_length": 8, "min_completion_length": 8, "max_completion_length": 41, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.00047409328483865383, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.4768168926239014, "avg_step_time_sec": 18.625491758872723, "gpu_allocated_gb": 32.362806272, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2162.4878602027893, "eta_sec": 1564.5413077453088, "timestamp": "2026-02-15T12:46:53.100381"}
{"step": 117, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.36, "mean_correctness_reward": 0.36, "mean_length_penalty": -0.008466796875, "mean_total_reward": 0.3557666015625, "std_total_reward": 0.4772711323445163, "total_tokens": 867, "mean_completion_length": 17.34, "median_completion_length": 10, "min_completion_length": 4, "max_completion_length": 85, "frac_has_think": 0.78, "frac_has_boxed": 0.02, "mean_reward_std_per_prompt": 0.19626255383297508, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.4758274555206299, "avg_step_time_sec": 18.478913431493645, "gpu_allocated_gb": 32.362812928, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2163.970665693283, "eta_sec": 1533.7498148139725, "timestamp": "2026-02-15T12:46:54.947556"}
{"step": 118, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.36, "mean_correctness_reward": 0.36, "mean_length_penalty": -0.031865234375, "mean_total_reward": 0.3440673828125, "std_total_reward": 0.4896543127376206, "total_tokens": 3263, "mean_completion_length": 65.26, "median_completion_length": 16, "min_completion_length": 5, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.02, "mean_reward_std_per_prompt": 0.29836295184574246, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.926435947418213, "avg_step_time_sec": 18.33863819857775, "gpu_allocated_gb": 32.362979328, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2165.9025032520294, "eta_sec": 1503.7683322833757, "timestamp": "2026-02-15T12:47:08.559764"}
{"step": 119, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.08, "mean_correctness_reward": 0.08, "mean_length_penalty": -0.005068359375, "mean_total_reward": 0.0774658203125, "std_total_reward": 0.27146559945367793, "total_tokens": 519, "mean_completion_length": 10.38, "median_completion_length": 10, "min_completion_length": 8, "max_completion_length": 18, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.09812186317113507, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 20.880955457687378, "avg_step_time_sec": 18.360002209158505, "gpu_allocated_gb": 32.365689856, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2186.7888231277466, "eta_sec": 1487.160178941839, "timestamp": "2026-02-15T12:47:17.272189"}
{"step": 120, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.00455078125, "mean_total_reward": 0.397724609375, "std_total_reward": 0.4899621113900945, "total_tokens": 466, "mean_completion_length": 9.32, "median_completion_length": 8, "min_completion_length": 8, "max_completion_length": 22, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0002443358594374376, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.3822047710418701, "avg_step_time_sec": 18.218520563840865, "gpu_allocated_gb": 32.362801152, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2188.1769568920135, "eta_sec": 1457.4816451072693, "timestamp": "2026-02-15T12:47:18.708055"}
{"step": 121, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.14, "mean_correctness_reward": 0.14, "mean_length_penalty": -0.004931640625, "mean_total_reward": 0.1375341796875, "std_total_reward": 0.34680008503699444, "total_tokens": 505, "mean_completion_length": 10.1, "median_completion_length": 10, "min_completion_length": 8, "max_completion_length": 12, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.09170154783840372, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.4449543952941895, "avg_step_time_sec": 18.079896050051225, "gpu_allocated_gb": 32.36284672, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2189.6272859573364, "eta_sec": 1428.3117879540468, "timestamp": "2026-02-15T12:47:20.102864"}
{"step": 122, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.005615234375, "mean_total_reward": -0.0028076171875, "std_total_reward": 0.0007552591942508799, "total_tokens": 575, "mean_completion_length": 11.5, "median_completion_length": 10, "min_completion_length": 5, "max_completion_length": 17, "frac_has_think": 0.9, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.00027069867634075086, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3832955360412598, "avg_step_time_sec": 17.943038668788848, "gpu_allocated_gb": 32.362819072, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2191.0176758766174, "eta_sec": 1399.55701616553, "timestamp": "2026-02-15T12:47:21.551118"}
{"step": 123, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.46, "mean_correctness_reward": 0.46, "mean_length_penalty": -0.005302734375, "mean_total_reward": 0.4573486328125, "std_total_reward": 0.4978277039825287, "total_tokens": 543, "mean_completion_length": 10.86, "median_completion_length": 10, "min_completion_length": 5, "max_completion_length": 20, "frac_has_think": 0.98, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.15798039611459816, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.4256024360656738, "avg_step_time_sec": 17.808750569335814, "gpu_allocated_gb": 32.36280064, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2192.4516377449036, "eta_sec": 1371.2737938388577, "timestamp": "2026-02-15T12:47:22.952287"}
{"step": 124, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.0255859375, "mean_total_reward": 0.38720703125, "std_total_reward": 0.5019372413586138, "total_tokens": 2620, "mean_completion_length": 52.4, "median_completion_length": 10, "min_completion_length": 5, "max_completion_length": 2048, "frac_has_think": 0.7, "frac_has_boxed": 0.02, "mean_reward_std_per_prompt": 0.030931206146740418, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.3911628723144531, "avg_step_time_sec": 17.676350668553383, "gpu_allocated_gb": 32.362785792, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2193.848342180252, "eta_sec": 1343.4026508100571, "timestamp": "2026-02-15T12:47:36.350137"}
{"step": 125, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.06, "mean_correctness_reward": 0.06, "mean_length_penalty": -0.00498046875, "mean_total_reward": 0.057509765625, "std_total_reward": 0.2374587633573386, "total_tokens": 510, "mean_completion_length": 10.2, "median_completion_length": 10, "min_completion_length": 8, "max_completion_length": 14, "frac_has_think": 0.94, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.18002117409515123, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 20.698216915130615, "avg_step_time_sec": 17.700525598526, "gpu_allocated_gb": 32.365631488, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2214.5518260002136, "eta_sec": 1327.5394198894498, "timestamp": "2026-02-15T12:47:45.086570"}
{"step": 126, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.005625, "mean_total_reward": -0.0028125, "std_total_reward": 0.0008751547540325993, "total_tokens": 576, "mean_completion_length": 11.52, "median_completion_length": 10, "min_completion_length": 6, "max_completion_length": 18, "frac_has_think": 0.72, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.00019235291461620186, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.436293363571167, "avg_step_time_sec": 17.571444390312074, "gpu_allocated_gb": 32.36284928, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2215.9939715862274, "eta_sec": 1300.2868848830935, "timestamp": "2026-02-15T12:47:46.515009"}
{"step": 127, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.06, "mean_correctness_reward": 0.06, "mean_length_penalty": -0.003525390625, "mean_total_reward": 0.0582373046875, "std_total_reward": 0.23731579016312804, "total_tokens": 361, "mean_completion_length": 7.22, "median_completion_length": 6, "min_completion_length": 4, "max_completion_length": 10, "frac_has_think": 0.28, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.09168898660321168, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.4164433479309082, "avg_step_time_sec": 17.4442396576949, "gpu_allocated_gb": 32.36279808, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2217.41868185997, "eta_sec": 1273.4294950117278, "timestamp": "2026-02-15T12:47:47.889337"}
{"step": 128, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00583984375, "mean_total_reward": -0.002919921875, "std_total_reward": 0.0013183232055223597, "total_tokens": 598, "mean_completion_length": 11.96, "median_completion_length": 10, "min_completion_length": 8, "max_completion_length": 29, "frac_has_think": 0.72, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.000450126802403749, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3504157066345215, "avg_step_time_sec": 17.31850665807724, "gpu_allocated_gb": 32.362770432, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2218.7777049541473, "eta_sec": 1246.9324793815613, "timestamp": "2026-02-15T12:47:49.334417"}
{"step": 129, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.38, "mean_correctness_reward": 0.38, "mean_length_penalty": -0.00349609375, "mean_total_reward": 0.378251953125, "std_total_reward": 0.4851154585197091, "total_tokens": 358, "mean_completion_length": 7.16, "median_completion_length": 8, "min_completion_length": 4, "max_completion_length": 10, "frac_has_think": 0.38, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.06030361651459625, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.4501917362213135, "avg_step_time_sec": 17.19549646488456, "gpu_allocated_gb": 32.362820608, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2220.233367204666, "eta_sec": 1220.8802490068038, "timestamp": "2026-02-15T12:47:50.676392"}
{"step": 130, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.004658203125, "mean_total_reward": -0.0023291015625, "std_total_reward": 0.00320716045458557, "total_tokens": 477, "mean_completion_length": 9.54, "median_completion_length": 4, "min_completion_length": 4, "max_completion_length": 62, "frac_has_think": 0.02, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0009819841759181674, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3381097316741943, "avg_step_time_sec": 17.073516566936785, "gpu_allocated_gb": 32.36278784, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2221.576704978943, "eta_sec": 1195.146159685575, "timestamp": "2026-02-15T12:47:52.321710"}
{"step": 131, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.02, "mean_correctness_reward": 0.02, "mean_length_penalty": -0.023046875, "mean_total_reward": 0.0084765625, "std_total_reward": 0.15767868744321764, "total_tokens": 2360, "mean_completion_length": 47.2, "median_completion_length": 5, "min_completion_length": 4, "max_completion_length": 2048, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.08988317717579145, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.6358296871185303, "avg_step_time_sec": 16.955671628922907, "gpu_allocated_gb": 32.362864128, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2223.2179832458496, "eta_sec": 1169.9413423956805, "timestamp": "2026-02-15T12:48:05.628839"}
{"step": 132, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.00205078125, "mean_total_reward": 0.198974609375, "std_total_reward": 0.40014649927070595, "total_tokens": 210, "mean_completion_length": 4.2, "median_completion_length": 4, "min_completion_length": 3, "max_completion_length": 5, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 20.19326400756836, "avg_step_time_sec": 16.980198843912646, "gpu_allocated_gb": 32.365605888, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2243.4163224697113, "eta_sec": 1154.65352138606, "timestamp": "2026-02-15T12:48:13.811480"}
{"step": 133, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.00162109375, "mean_total_reward": 0.199189453125, "std_total_reward": 0.4001611582832244, "total_tokens": 166, "mean_completion_length": 3.32, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 6, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 7.404859077105312e-05, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.2741918563842773, "avg_step_time_sec": 16.862108565810928, "gpu_allocated_gb": 32.362788864, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2244.695714712143, "eta_sec": 1129.7612739093322, "timestamp": "2026-02-15T12:48:15.116084"}
{"step": 134, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00224609375, "mean_total_reward": -0.001123046875, "std_total_reward": 0.00047841596538733945, "total_tokens": 230, "mean_completion_length": 4.6, "median_completion_length": 5, "min_completion_length": 2, "max_completion_length": 12, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.00011960399134683486, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3105359077453613, "avg_step_time_sec": 16.746052053437303, "gpu_allocated_gb": 32.362784256, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2246.0132822990417, "eta_sec": 1105.239435526862, "timestamp": "2026-02-15T12:48:16.494219"}
{"step": 135, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.22, "mean_correctness_reward": 0.22, "mean_length_penalty": -0.003134765625, "mean_total_reward": 0.2184326171875, "std_total_reward": 0.41457505153965546, "total_tokens": 321, "mean_completion_length": 6.42, "median_completion_length": 4, "min_completion_length": 2, "max_completion_length": 107, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.061523604428406244, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.3817355632781982, "avg_step_time_sec": 16.632242301658348, "gpu_allocated_gb": 32.362807808, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2247.4006068706512, "eta_sec": 1081.0957496077926, "timestamp": "2026-02-15T12:48:18.343513"}
{"step": 136, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.36, "mean_correctness_reward": 0.36, "mean_length_penalty": -0.00328125, "mean_total_reward": 0.358359375, "std_total_reward": 0.4803779339071667, "total_tokens": 336, "mean_completion_length": 6.72, "median_completion_length": 5, "min_completion_length": 2, "max_completion_length": 26, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.17848942720579317, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.8614349365234375, "avg_step_time_sec": 16.523633423973532, "gpu_allocated_gb": 32.362898432, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2249.2675564289093, "eta_sec": 1057.512539134306, "timestamp": "2026-02-15T12:48:19.803649"}
{"step": 137, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.16, "mean_correctness_reward": 0.16, "mean_length_penalty": -0.00248046875, "mean_total_reward": 0.158759765625, "std_total_reward": 0.36574935767533906, "total_tokens": 254, "mean_completion_length": 5.08, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 15, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.08002685751697206, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.4228260517120361, "avg_step_time_sec": 16.413408552643155, "gpu_allocated_gb": 32.36280576, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2250.6971802711487, "eta_sec": 1034.0447388165187, "timestamp": "2026-02-15T12:48:21.163047"}
{"step": 138, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.22, "mean_correctness_reward": 0.22, "mean_length_penalty": -0.001220703125, "mean_total_reward": 0.2193896484375, "std_total_reward": 0.4140636138613132, "total_tokens": 125, "mean_completion_length": 2.5, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 4, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0600244140625, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.7931873798370361, "avg_step_time_sec": 16.307464920956157, "gpu_allocated_gb": 32.362800128, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2252.4953849315643, "eta_sec": 1011.0628250992818, "timestamp": "2026-02-15T12:48:22.881596"}
{"step": 139, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.02, "mean_correctness_reward": 0.02, "mean_length_penalty": -0.001728515625, "mean_total_reward": 0.0191357421875, "std_total_reward": 0.14001905558533662, "total_tokens": 177, "mean_completion_length": 3.54, "median_completion_length": 4, "min_completion_length": 2, "max_completion_length": 5, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.06001139631919104, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.2853896617889404, "avg_step_time_sec": 16.199392437077254, "gpu_allocated_gb": 32.362790912, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2253.786205768585, "eta_sec": 988.1629386617125, "timestamp": "2026-02-15T12:48:24.205961"}
{"step": 140, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.001669921875, "mean_total_reward": 0.3991650390625, "std_total_reward": 0.48978233254139686, "total_tokens": 171, "mean_completion_length": 3.42, "median_completion_length": 3, "min_completion_length": 3, "max_completion_length": 4, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 1.46484375e-05, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.319098949432373, "avg_step_time_sec": 16.09310462645122, "gpu_allocated_gb": 32.36281856, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2255.1121575832367, "eta_sec": 965.5862775870731, "timestamp": "2026-02-15T12:48:25.501454"}
{"step": 141, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.001552734375, "mean_total_reward": -0.0007763671875, "std_total_reward": 0.00027917563201021103, "total_tokens": 159, "mean_completion_length": 3.18, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 5, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 1.46484375e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.2731938362121582, "avg_step_time_sec": 15.987998876165836, "gpu_allocated_gb": 32.362783744, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2256.391512155533, "eta_sec": 943.2919336937844, "timestamp": "2026-02-15T12:48:26.788673"}
{"step": 142, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.00201171875, "mean_total_reward": 0.498994140625, "std_total_reward": 0.4996391177536588, "total_tokens": 206, "mean_completion_length": 4.12, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 12, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.15798496578805776, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 1.3009722232818604, "avg_step_time_sec": 15.884569111004682, "gpu_allocated_gb": 32.3627776, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2257.6974923610687, "eta_sec": 921.3050084382716, "timestamp": "2026-02-15T12:48:28.143421"}
{"step": 143, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.0021484375, "mean_total_reward": 0.19892578125, "std_total_reward": 0.400293516711018, "total_tokens": 220, "mean_completion_length": 4.4, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 12, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.00021337337149686728, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.3414804935455322, "avg_step_time_sec": 15.78286919060287, "gpu_allocated_gb": 32.362774016, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2259.0452032089233, "eta_sec": 899.6235438643636, "timestamp": "2026-02-15T12:48:29.492773"}
{"step": 144, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.24, "mean_correctness_reward": 0.24, "mean_length_penalty": -0.001962890625, "mean_total_reward": 0.2390185546875, "std_total_reward": 0.42729176231927546, "total_tokens": 201, "mean_completion_length": 4.02, "median_completion_length": 4, "min_completion_length": 2, "max_completion_length": 7, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0800146484375, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.3376080989837646, "avg_step_time_sec": 15.682554877466625, "gpu_allocated_gb": 32.362798592, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2260.390387058258, "eta_sec": 878.223073138131, "timestamp": "2026-02-15T12:48:30.803876"}
{"step": 145, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.00109375, "mean_total_reward": 0.399453125, "std_total_reward": 0.4898461285157689, "total_tokens": 112, "mean_completion_length": 2.24, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 1.953125e-05, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.3128368854522705, "avg_step_time_sec": 15.583453374073423, "gpu_allocated_gb": 32.362772992, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2261.708656311035, "eta_sec": 857.0899355740382, "timestamp": "2026-02-15T12:48:32.101507"}
{"step": 146, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.0015625, "mean_total_reward": 0.39921875, "std_total_reward": 0.48953935574451407, "total_tokens": 160, "mean_completion_length": 3.2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 8, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.2953195571899414, "avg_step_time_sec": 15.485589443820796, "gpu_allocated_gb": 32.362796032, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2263.010260820389, "eta_sec": 836.221829966323, "timestamp": "2026-02-15T12:48:33.446241"}
{"step": 147, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.16, "mean_correctness_reward": 0.16, "mean_length_penalty": -0.00126953125, "mean_total_reward": 0.159365234375, "std_total_reward": 0.36656345194947454, "total_tokens": 130, "mean_completion_length": 2.6, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.08, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.3348357677459717, "avg_step_time_sec": 15.389325813371308, "gpu_allocated_gb": 32.362784768, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2264.3542790412903, "eta_sec": 815.6342681086793, "timestamp": "2026-02-15T12:48:34.746144"}
{"step": 148, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.22, "mean_correctness_reward": 0.22, "mean_length_penalty": -0.004775390625, "mean_total_reward": 0.2176123046875, "std_total_reward": 0.4152340954037691, "total_tokens": 489, "mean_completion_length": 9.78, "median_completion_length": 4, "min_completion_length": 2, "max_completion_length": 57, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0609460826122754, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.2807726860046387, "avg_step_time_sec": 15.293997751699912, "gpu_allocated_gb": 32.362783744, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2265.6406784057617, "eta_sec": 795.2878830883955, "timestamp": "2026-02-15T12:48:36.379227"}
{"step": 149, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.00140625, "mean_total_reward": 0.579296875, "std_total_reward": 0.4935142388806693, "total_tokens": 144, "mean_completion_length": 2.88, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 5, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.06009028318357811, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.646324634552002, "avg_step_time_sec": 15.202402630108315, "gpu_allocated_gb": 32.362871296, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2267.292567253113, "eta_sec": 775.322534135524, "timestamp": "2026-02-15T12:48:37.689594"}
{"step": 150, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.36, "mean_correctness_reward": 0.36, "mean_length_penalty": -0.001513671875, "mean_total_reward": 0.3592431640625, "std_total_reward": 0.4797946256374341, "total_tokens": 155, "mean_completion_length": 3.1, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 6, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.08002441406250001, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.6613857746124268, "avg_step_time_sec": 15.112129184405008, "gpu_allocated_gb": 32.362743808, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2268.9595510959625, "eta_sec": 755.6064592202504, "timestamp": "2026-02-15T12:48:39.993985"}
{"step": 151, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.00146484375, "mean_total_reward": 0.599267578125, "std_total_reward": 0.4897983414381614, "total_tokens": 150, "mean_completion_length": 3, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 5, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.3140878677368164, "avg_step_time_sec": 15.020751427340981, "gpu_allocated_gb": 32.36277504, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2270.9105355739594, "eta_sec": 736.0168199397081, "timestamp": "2026-02-15T12:48:41.358951"}
{"step": 152, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.12, "mean_correctness_reward": 0.12, "mean_length_penalty": -0.001474609375, "mean_total_reward": 0.1192626953125, "std_total_reward": 0.3250536260011521, "total_tokens": 151, "mean_completion_length": 3.02, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 5, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.09799423814882711, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.362705945968628, "avg_step_time_sec": 14.930895864963531, "gpu_allocated_gb": 32.362835968, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2272.2812769412994, "eta_sec": 716.6830015182495, "timestamp": "2026-02-15T12:48:42.689682"}
{"step": 153, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.001494140625, "mean_total_reward": -0.0007470703125, "std_total_reward": 0.000456510328462211, "total_tokens": 153, "mean_completion_length": 3.06, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 8, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 7.674855674128849e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3092501163482666, "avg_step_time_sec": 14.841865500593498, "gpu_allocated_gb": 32.362785792, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2273.596509695053, "eta_sec": 697.5676785278944, "timestamp": "2026-02-15T12:48:44.030122"}
{"step": 154, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.0016796875, "mean_total_reward": 0.19916015625, "std_total_reward": 0.40005375548598765, "total_tokens": 172, "mean_completion_length": 3.44, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 6, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 7.8125e-05, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.3388919830322266, "avg_step_time_sec": 14.754183854375567, "gpu_allocated_gb": 32.362834944, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2274.9409449100494, "eta_sec": 678.6924573012761, "timestamp": "2026-02-15T12:48:45.342573"}
{"step": 155, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00107421875, "mean_total_reward": -0.000537109375, "std_total_reward": 9.765625e-05, "total_tokens": 110, "mean_completion_length": 2.2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3283231258392334, "avg_step_time_sec": 14.667565398062429, "gpu_allocated_gb": 32.362789376, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2276.274568796158, "eta_sec": 660.0404429128093, "timestamp": "2026-02-15T12:48:46.669090"}
{"step": 156, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.28, "mean_correctness_reward": 0.28, "mean_length_penalty": -0.001630859375, "mean_total_reward": 0.2791845703125, "std_total_reward": 0.44878986141755617, "total_tokens": 167, "mean_completion_length": 3.34, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 11, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.15145013117814707, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.3037376403808594, "avg_step_time_sec": 14.581899835513187, "gpu_allocated_gb": 32.362807296, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2277.58380150795, "eta_sec": 641.6035927625802, "timestamp": "2026-02-15T12:48:48.033082"}
{"step": 157, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.02, "mean_correctness_reward": 0.02, "mean_length_penalty": -0.001015625, "mean_total_reward": 0.0194921875, "std_total_reward": 0.13986328125, "total_tokens": 104, "mean_completion_length": 2.08, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 6, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.059941406249999996, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.3514611721038818, "avg_step_time_sec": 14.497629525555167, "gpu_allocated_gb": 32.362798592, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2278.943127632141, "eta_sec": 623.3980695988722, "timestamp": "2026-02-15T12:48:49.346323"}
{"step": 158, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.12, "mean_correctness_reward": 0.12, "mean_length_penalty": -0.001708984375, "mean_total_reward": 0.1191455078125, "std_total_reward": 0.3246461206747534, "total_tokens": 175, "mean_completion_length": 3.5, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 7, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.09801475592728838, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.30765700340271, "avg_step_time_sec": 14.414148686807367, "gpu_allocated_gb": 32.362823168, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2280.256495475769, "eta_sec": 605.3942448459094, "timestamp": "2026-02-15T12:48:50.676292"}
{"step": 159, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00103515625, "mean_total_reward": -0.000517578125, "std_total_reward": 7.933631254527306e-05, "total_tokens": 106, "mean_completion_length": 2.12, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 2.392079826936697e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.351884365081787, "avg_step_time_sec": 14.33199608101035, "gpu_allocated_gb": 32.362818048, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2281.614112854004, "eta_sec": 587.6118393214243, "timestamp": "2026-02-15T12:48:51.996240"}
{"step": 160, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.24, "mean_correctness_reward": 0.24, "mean_length_penalty": -0.00107421875, "mean_total_reward": 0.239462890625, "std_total_reward": 0.42711057940156805, "total_tokens": 110, "mean_completion_length": 2.2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.08, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.2799596786499023, "avg_step_time_sec": 14.250420853495598, "gpu_allocated_gb": 32.362784768, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2282.8993458747864, "eta_sec": 570.0168341398239, "timestamp": "2026-02-15T12:48:53.294537"}
{"step": 161, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.0015234375, "mean_total_reward": -0.00076171875, "std_total_reward": 0.0003642198476361624, "total_tokens": 156, "mean_completion_length": 3.12, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 6, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 7.713847153873394e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3020579814910889, "avg_step_time_sec": 14.169996239383769, "gpu_allocated_gb": 32.362822656, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2284.2070479393005, "eta_sec": 552.629853335967, "timestamp": "2026-02-15T12:48:54.628868"}
{"step": 162, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.02, "mean_correctness_reward": 0.02, "mean_length_penalty": -0.00138671875, "mean_total_reward": 0.019306640625, "std_total_reward": 0.1399599229124805, "total_tokens": 142, "mean_completion_length": 2.84, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 9, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.060197622827668815, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.7617032527923584, "avg_step_time_sec": 14.09340183823197, "gpu_allocated_gb": 32.362760704, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2285.9751374721527, "eta_sec": 535.5492698528149, "timestamp": "2026-02-15T12:48:56.419471"}
{"step": 163, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.002265625, "mean_total_reward": -0.0011328125, "std_total_reward": 0.0015714465548984935, "total_tokens": 232, "mean_completion_length": 4.64, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 43, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.00046875, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3595497608184814, "avg_step_time_sec": 14.015280046345998, "gpu_allocated_gb": 32.362807296, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2287.342102766037, "eta_sec": 518.565361714802, "timestamp": "2026-02-15T12:48:57.951937"}
{"step": 164, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.34, "mean_correctness_reward": 0.34, "mean_length_penalty": -0.003251953125, "mean_total_reward": 0.3383740234375, "std_total_reward": 0.4729287425808672, "total_tokens": 333, "mean_completion_length": 6.66, "median_completion_length": 4, "min_completion_length": 3, "max_completion_length": 21, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.1401642406058715, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.5218725204467773, "avg_step_time_sec": 13.939100732163686, "gpu_allocated_gb": 32.362819072, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2288.870108127594, "eta_sec": 501.8076263578927, "timestamp": "2026-02-15T12:48:59.404465"}
{"step": 165, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.0015625, "mean_total_reward": 0.19921875, "std_total_reward": 0.3996582478667125, "total_tokens": 160, "mean_completion_length": 3.2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 6, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.4453423023223877, "avg_step_time_sec": 13.86338098410404, "gpu_allocated_gb": 32.362845696, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2290.322884082794, "eta_sec": 485.2183344436414, "timestamp": "2026-02-15T12:49:00.741934"}
{"step": 166, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.002158203125, "mean_total_reward": -0.0010791015625, "std_total_reward": 0.0018224668401937954, "total_tokens": 221, "mean_completion_length": 4.42, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 49, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0007599134386069888, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3392350673675537, "avg_step_time_sec": 13.787934321955026, "gpu_allocated_gb": 32.362765824, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2291.6696496009827, "eta_sec": 468.7897669464709, "timestamp": "2026-02-15T12:49:02.360430"}
{"step": 167, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.0029296875, "mean_total_reward": 0.55853515625, "std_total_reward": 0.4960750271365441, "total_tokens": 300, "mean_completion_length": 6, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 20, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.08, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.6087417602539062, "avg_step_time_sec": 13.715005025178373, "gpu_allocated_gb": 32.36288512, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2293.2852749824524, "eta_sec": 452.5951658308863, "timestamp": "2026-02-15T12:49:03.790095"}
{"step": 168, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00123046875, "mean_total_reward": -0.000615234375, "std_total_reward": 0.00017082866879429592, "total_tokens": 126, "mean_completion_length": 2.52, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 6, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 5.859375e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.4152333736419678, "avg_step_time_sec": 13.641792098681131, "gpu_allocated_gb": 32.36283904, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2294.706067085266, "eta_sec": 436.5373471577962, "timestamp": "2026-02-15T12:49:05.121959"}
{"step": 169, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00205078125, "mean_total_reward": -0.001025390625, "std_total_reward": 0.0008910296675241532, "total_tokens": 210, "mean_completion_length": 4.2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 12, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 6.86725140404201e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3234431743621826, "avg_step_time_sec": 13.568902460075694, "gpu_allocated_gb": 32.362774016, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2296.036595106125, "eta_sec": 420.6359762623465, "timestamp": "2026-02-15T12:49:06.491890"}
{"step": 170, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.26, "mean_correctness_reward": 0.26, "mean_length_penalty": -0.00126953125, "mean_total_reward": 0.259365234375, "std_total_reward": 0.4387211616798174, "total_tokens": 130, "mean_completion_length": 2.6, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 5, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0916515138991168, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.3626937866210938, "avg_step_time_sec": 13.497101232584786, "gpu_allocated_gb": 32.362768896, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2297.407915353775, "eta_sec": 404.9130369775436, "timestamp": "2026-02-15T12:49:07.792856"}
{"step": 171, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.0021875, "mean_total_reward": 0.39890625, "std_total_reward": 0.4898941350938953, "total_tokens": 224, "mean_completion_length": 4.48, "median_completion_length": 6, "min_completion_length": 2, "max_completion_length": 6, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 3.238891396831445e-05, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.2797670364379883, "avg_step_time_sec": 13.425654833776909, "gpu_allocated_gb": 32.36277248, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2298.693243265152, "eta_sec": 389.34399017953035, "timestamp": "2026-02-15T12:49:09.129670"}
{"step": 172, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.42, "mean_correctness_reward": 0.42, "mean_length_penalty": -0.001318359375, "mean_total_reward": 0.4193408203125, "std_total_reward": 0.49360503927364163, "total_tokens": 135, "mean_completion_length": 2.7, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 4, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.06000814243539328, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 1.346221685409546, "avg_step_time_sec": 13.355425571286402, "gpu_allocated_gb": 32.362842624, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2300.044274330139, "eta_sec": 373.95191599601924, "timestamp": "2026-02-15T12:49:10.441342"}
{"step": 173, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.001279296875, "mean_total_reward": 0.1993603515625, "std_total_reward": 0.4000757237820263, "total_tokens": 131, "mean_completion_length": 2.62, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 4, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 1.46484375e-05, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.2935268878936768, "avg_step_time_sec": 13.285703613578928, "gpu_allocated_gb": 32.362779648, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2301.343810558319, "eta_sec": 358.71399756663106, "timestamp": "2026-02-15T12:49:11.727318"}
{"step": 174, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.001494140625, "mean_total_reward": 0.1992529296875, "std_total_reward": 0.40012954762838004, "total_tokens": 153, "mean_completion_length": 3.06, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 7, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 7.032775506559008e-05, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.7588646411895752, "avg_step_time_sec": 13.219457412588186, "gpu_allocated_gb": 32.362760192, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2303.108006477356, "eta_sec": 343.70589272729285, "timestamp": "2026-02-15T12:49:13.512741"}
{"step": 175, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.02, "mean_correctness_reward": 0.02, "mean_length_penalty": -0.001474609375, "mean_total_reward": 0.0192626953125, "std_total_reward": 0.13993127949950457, "total_tokens": 151, "mean_completion_length": 3.02, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 5, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0599853515625, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.3061587810516357, "avg_step_time_sec": 13.151381420407976, "gpu_allocated_gb": 32.362779136, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2304.4197974205017, "eta_sec": 328.7845355101994, "timestamp": "2026-02-15T12:49:14.823461"}
{"step": 176, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.001171875, "mean_total_reward": -0.0005859375, "std_total_reward": 0.0001953125, "total_tokens": 120, "mean_completion_length": 2.4, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 4, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.2869510650634766, "avg_step_time_sec": 13.083969884298064, "gpu_allocated_gb": 32.362778624, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2305.7150785923004, "eta_sec": 314.01527722315353, "timestamp": "2026-02-15T12:49:16.113551"}
{"step": 177, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.00126953125, "mean_total_reward": 0.399365234375, "std_total_reward": 0.4898182206748895, "total_tokens": 130, "mean_completion_length": 2.6, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.3020994663238525, "avg_step_time_sec": 13.017405644648493, "gpu_allocated_gb": 32.362755072, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2307.024134874344, "eta_sec": 299.4003298269153, "timestamp": "2026-02-15T12:49:17.418551"}
{"step": 178, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00130859375, "mean_total_reward": -0.000654296875, "std_total_reward": 0.00019797983332727826, "total_tokens": 134, "mean_completion_length": 2.68, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 4, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 3.702429538552656e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.2866466045379639, "avg_step_time_sec": 12.951502503973714, "gpu_allocated_gb": 32.362781696, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2308.3176181316376, "eta_sec": 284.93305508742174, "timestamp": "2026-02-15T12:49:18.727924"}
{"step": 179, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.1, "mean_correctness_reward": 0.1, "mean_length_penalty": -0.003076171875, "mean_total_reward": 0.0984619140625, "std_total_reward": 0.2998725352104504, "total_tokens": 315, "mean_completion_length": 6.3, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 58, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.10075697411095993, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.2860527038574219, "avg_step_time_sec": 12.88633239335854, "gpu_allocated_gb": 32.362785792, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2309.611048221588, "eta_sec": 270.61298026052935, "timestamp": "2026-02-15T12:49:20.414444"}
{"step": 180, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00107421875, "mean_total_reward": -0.000537109375, "std_total_reward": 9.765625e-05, "total_tokens": 110, "mean_completion_length": 2.2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.723142385482788, "avg_step_time_sec": 12.824314671092564, "gpu_allocated_gb": 32.36291328, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2311.3431346416473, "eta_sec": 256.48629342185126, "timestamp": "2026-02-15T12:49:21.721971"}
{"step": 181, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.18, "mean_correctness_reward": 0.18, "mean_length_penalty": -0.00140625, "mean_total_reward": 0.179296875, "std_total_reward": 0.38417386630932315, "total_tokens": 144, "mean_completion_length": 2.88, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 6, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.06002392079826936, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.2752256393432617, "avg_step_time_sec": 12.76050754937019, "gpu_allocated_gb": 32.36279296, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2312.6243889331818, "eta_sec": 242.44964343803363, "timestamp": "2026-02-15T12:49:23.031174"}
{"step": 182, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.000986328125, "mean_total_reward": 0.1995068359375, "std_total_reward": 0.40000244285910436, "total_tokens": 101, "mean_completion_length": 2.02, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 1.46484375e-05, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.293766975402832, "avg_step_time_sec": 12.697503480282458, "gpu_allocated_gb": 32.362772992, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2313.925259590149, "eta_sec": 228.55506264508426, "timestamp": "2026-02-15T12:49:24.304834"}
{"step": 183, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.04, "mean_correctness_reward": 0.04, "mean_length_penalty": -0.001455078125, "mean_total_reward": 0.0392724609375, "std_total_reward": 0.19533541741767446, "total_tokens": 149, "mean_completion_length": 2.98, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 20, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0797949824094163, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.2947609424591064, "avg_step_time_sec": 12.635193411769762, "gpu_allocated_gb": 32.362758144, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2315.225583076477, "eta_sec": 214.79828800008596, "timestamp": "2026-02-15T12:49:25.712398"}
{"step": 184, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.0017578125, "mean_total_reward": -0.00087890625, "std_total_reward": 0.000603970550627587, "total_tokens": 180, "mean_completion_length": 3.6, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 13, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0002509891384462765, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3810811042785645, "avg_step_time_sec": 12.574029757924702, "gpu_allocated_gb": 32.362797056, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2316.615049123764, "eta_sec": 201.18447612679523, "timestamp": "2026-02-15T12:49:27.084472"}
{"step": 185, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.00146484375, "mean_total_reward": 0.199267578125, "std_total_reward": 0.39970709089832807, "total_tokens": 150, "mean_completion_length": 3, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 8, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.00012215904037655534, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.3484950065612793, "avg_step_time_sec": 12.513351191701116, "gpu_allocated_gb": 32.362817536, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2317.972700357437, "eta_sec": 187.70026787551674, "timestamp": "2026-02-15T12:49:28.419151"}
{"step": 186, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.0012109375, "mean_total_reward": -0.00060546875, "std_total_reward": 0.00015625, "total_tokens": 124, "mean_completion_length": 2.48, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 4, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 2.392079826936697e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.7716724872589111, "avg_step_time_sec": 12.455600230924544, "gpu_allocated_gb": 32.362841088, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2319.750196456909, "eta_sec": 174.3784032329436, "timestamp": "2026-02-15T12:49:30.142096"}
{"step": 187, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.32, "mean_correctness_reward": 0.32, "mean_length_penalty": -0.00126953125, "mean_total_reward": 0.319365234375, "std_total_reward": 0.46640917047707664, "total_tokens": 130, "mean_completion_length": 2.6, "median_completion_length": 3, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.09797958971132711, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.300215482711792, "avg_step_time_sec": 12.395945767030359, "gpu_allocated_gb": 32.36279296, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2321.056284427643, "eta_sec": 161.14729497139467, "timestamp": "2026-02-15T12:49:31.490921"}
{"step": 188, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.001357421875, "mean_total_reward": -0.0006787109375, "std_total_reward": 0.0002455525578257843, "total_tokens": 139, "mean_completion_length": 2.78, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 5, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 3.4179687499999996e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.3409478664398193, "avg_step_time_sec": 12.337142586708069, "gpu_allocated_gb": 32.362845184, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2322.4045808315277, "eta_sec": 148.04571104049683, "timestamp": "2026-02-15T12:49:32.823664"}
{"step": 189, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.001171875, "mean_total_reward": 0.1994140625, "std_total_reward": 0.3998046875, "total_tokens": 120, "mean_completion_length": 2.4, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 4, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.3297481536865234, "avg_step_time_sec": 12.27890240452277, "gpu_allocated_gb": 32.362797056, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2323.7407224178314, "eta_sec": 135.06792644975047, "timestamp": "2026-02-15T12:49:34.122359"}
{"step": 190, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.1, "mean_correctness_reward": 0.1, "mean_length_penalty": -0.00107421875, "mean_total_reward": 0.099462890625, "std_total_reward": 0.300016291493884, "total_tokens": 110, "mean_completion_length": 2.2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.1, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.2766876220703125, "avg_step_time_sec": 12.220996010930914, "gpu_allocated_gb": 32.362758144, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2325.0238020420074, "eta_sec": 122.20996010930915, "timestamp": "2026-02-15T12:49:35.413323"}
{"step": 191, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.001171875, "mean_total_reward": -0.0005859375, "std_total_reward": 0.0001953125, "total_tokens": 120, "mean_completion_length": 2.4, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 4, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.2721152305603027, "avg_step_time_sec": 12.163672027787612, "gpu_allocated_gb": 32.3627776, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2326.3039197921753, "eta_sec": 109.47304825008851, "timestamp": "2026-02-15T12:49:36.822725"}
{"step": 192, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.001142578125, "mean_total_reward": 0.3994287109375, "std_total_reward": 0.4897963174734083, "total_tokens": 117, "mean_completion_length": 2.34, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 6, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 5.371093750000001e-05, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.4102215766906738, "avg_step_time_sec": 12.107664473354816, "gpu_allocated_gb": 32.362891264, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2327.7239310741425, "eta_sec": 96.86131578683853, "timestamp": "2026-02-15T12:49:38.119268"}
{"step": 193, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00107421875, "mean_total_reward": -0.000537109375, "std_total_reward": 9.765625e-05, "total_tokens": 110, "mean_completion_length": 2.2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.2889833450317383, "avg_step_time_sec": 12.051609130721019, "gpu_allocated_gb": 32.362765824, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2329.018764257431, "eta_sec": 84.36126391504713, "timestamp": "2026-02-15T12:49:39.453959"}
{"step": 194, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00109375, "mean_total_reward": -0.000546875, "std_total_reward": 0.00011513502072804293, "total_tokens": 112, "mean_completion_length": 2.24, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 4, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 2.9296875e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.339081048965454, "avg_step_time_sec": 11.996389913804752, "gpu_allocated_gb": 32.362851328, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2330.364636182785, "eta_sec": 71.97833948282852, "timestamp": "2026-02-15T12:49:40.745964"}
{"step": 195, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.0009765625, "mean_total_reward": -0.00048828125, "std_total_reward": 0.0, "total_tokens": 100, "mean_completion_length": 2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 2, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.28251314163208, "avg_step_time_sec": 11.941446955998739, "gpu_allocated_gb": 32.362784768, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2331.6527075767517, "eta_sec": 59.7072347799937, "timestamp": "2026-02-15T12:49:42.027958"}
{"step": 196, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.0025, "mean_total_reward": -0.00125, "std_total_reward": 0.0015267204710517626, "total_tokens": 256, "mean_completion_length": 5.12, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 19, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 4.4751715771053124e-05, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.2790594100952148, "avg_step_time_sec": 11.887047019540047, "gpu_allocated_gb": 32.362763776, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2332.9372837543488, "eta_sec": 47.54818807816019, "timestamp": "2026-02-15T12:49:43.430493"}
{"step": 197, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.0009765625, "mean_total_reward": 0.19951171875, "std_total_reward": 0.4, "total_tokens": 100, "mean_completion_length": 2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 2, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 1.388758659362793, "avg_step_time_sec": 11.83375621568128, "gpu_allocated_gb": 32.362806272, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2334.331536769867, "eta_sec": 35.501268647043844, "timestamp": "2026-02-15T12:49:44.877222"}
{"step": 198, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00107421875, "mean_total_reward": -0.000537109375, "std_total_reward": 9.765625e-05, "total_tokens": 110, "mean_completion_length": 2.2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.524712085723877, "avg_step_time_sec": 11.78169033623705, "gpu_allocated_gb": 32.362961408, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2335.862363100052, "eta_sec": 23.5633806724741, "timestamp": "2026-02-15T12:49:46.241204"}
{"step": 199, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.0, "mean_correctness_reward": 0.0, "mean_length_penalty": -0.00107421875, "mean_total_reward": -0.000537109375, "std_total_reward": 9.765625e-05, "total_tokens": 110, "mean_completion_length": 2.2, "median_completion_length": 2, "min_completion_length": 2, "max_completion_length": 3, "frac_has_think": 0.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.0, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 1.0, "step_time_sec": 1.6977853775024414, "avg_step_time_sec": 11.731017446997178, "gpu_allocated_gb": 32.362769408, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2337.566358566284, "eta_sec": 11.731017446997178, "timestamp": "2026-02-15T12:49:48.000783"}