deepscaleR / gdpo_lambda0.1 /logs /step_stats.jsonl
brikdavies's picture
Add gdpo_lambda0.1/logs
f74c6a3 verified
Raw
History Blame Contribute Delete
163 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": 0.51839453125, "std_total_reward": 0.505042661530509, "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": 0.20587252117649119, "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": "2026-02-15T13:34:10.118263"}
{"step": 1, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.22, "mean_correctness_reward": 0.22, "mean_length_penalty": -0.94798828125, "mean_total_reward": 0.125201171875, "std_total_reward": 0.4209086011769524, "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": 0.26190294418025073, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 28.02942419052124, "avg_step_time_sec": 28.02942419052124, "gpu_allocated_gb": 32.365621248, "gpu_reserved_gb": 79.92246272, "gpu_peak_gb": 73.906566656, "elapsed_sec": 28.034409523010254, "eta_sec": 5577.855413913727, "timestamp": "2026-02-15T13:34:39.779655"}
{"step": 2, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.82720703125, "mean_total_reward": 0.517279296875, "std_total_reward": 0.5042072840400893, "total_tokens": 84706, "mean_completion_length": 1694.12, "median_completion_length": 1939, "min_completion_length": 809, "max_completion_length": 2048, "frac_has_think": 0.74, "frac_has_boxed": 0.74, "mean_reward_std_per_prompt": 0.4462199521814131, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 29.616747617721558, "avg_step_time_sec": 28.8230859041214, "gpu_allocated_gb": 32.365654016, "gpu_reserved_gb": 79.95392, "gpu_peak_gb": 74.127090176, "elapsed_sec": 57.658801794052124, "eta_sec": 5706.971009016037, "timestamp": "2026-02-15T13:35:07.852087"}
{"step": 3, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.751865234375, "mean_total_reward": 0.7248134765625, "std_total_reward": 0.4127584745402409, "total_tokens": 76991, "mean_completion_length": 1539.82, "median_completion_length": 1569, "min_completion_length": 558, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.1713938852873411, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.983965158462524, "avg_step_time_sec": 28.543378988901775, "gpu_allocated_gb": 32.365626368, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.127090176, "elapsed_sec": 85.64932322502136, "eta_sec": 5623.04566081365, "timestamp": "2026-02-15T13:35:35.756637"}
{"step": 4, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.24, "mean_correctness_reward": 0.24, "mean_length_penalty": -0.942685546875, "mean_total_reward": 0.14573144531249999, "std_total_reward": 0.4373162099450002, "total_tokens": 96531, "mean_completion_length": 1930.62, "median_completion_length": 2048, "min_completion_length": 1210, "max_completion_length": 2048, "frac_has_think": 0.28, "frac_has_boxed": 0.28, "mean_reward_std_per_prompt": 0.1963821588433302, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 27.92634129524231, "avg_step_time_sec": 28.389119565486908, "gpu_allocated_gb": 32.365638656, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.127090176, "elapsed_sec": 113.58215093612671, "eta_sec": 5564.267434835434, "timestamp": "2026-02-15T13:36:04.817126"}
{"step": 5, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.8410546875, "mean_total_reward": 0.41589453125, "std_total_reward": 0.5160480242936508, "total_tokens": 86124, "mean_completion_length": 1722.48, "median_completion_length": 2048, "min_completion_length": 853, "max_completion_length": 2048, "frac_has_think": 0.6, "frac_has_boxed": 0.62, "mean_reward_std_per_prompt": 0.3587214546829844, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 29.655746698379517, "avg_step_time_sec": 28.64244499206543, "gpu_allocated_gb": 32.365667328, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.156304384, "elapsed_sec": 143.24408769607544, "eta_sec": 5585.276773452759, "timestamp": "2026-02-15T13:36:33.588889"}
{"step": 6, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.794375, "mean_total_reward": 0.6605624999999999, "std_total_reward": 0.4510346238871666, "total_tokens": 81344, "mean_completion_length": 1626.88, "median_completion_length": 1733, "min_completion_length": 904, "max_completion_length": 2048, "frac_has_think": 0.78, "frac_has_boxed": 0.78, "mean_reward_std_per_prompt": 0.18923848323048292, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 28.130523920059204, "avg_step_time_sec": 28.557124813397724, "gpu_allocated_gb": 32.365632512, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.156304384, "elapsed_sec": 171.38196969032288, "eta_sec": 5540.082213799158, "timestamp": "2026-02-15T13:37:01.308098"}
{"step": 7, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.46, "mean_correctness_reward": 0.46, "mean_length_penalty": -0.86125, "mean_total_reward": 0.373875, "std_total_reward": 0.5135556230552336, "total_tokens": 88192, "mean_completion_length": 1763.84, "median_completion_length": 2048, "min_completion_length": 783, "max_completion_length": 2048, "frac_has_think": 0.5, "frac_has_boxed": 0.5, "mean_reward_std_per_prompt": 0.24454921164323962, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 28.128169536590576, "avg_step_time_sec": 28.49584548813956, "gpu_allocated_gb": 32.36562432, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.156304384, "elapsed_sec": 199.51522302627563, "eta_sec": 5499.698179210935, "timestamp": "2026-02-15T13:37:30.003580"}
{"step": 8, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.791650390625, "mean_total_reward": 0.5408349609375, "std_total_reward": 0.4923794629150635, "total_tokens": 81065, "mean_completion_length": 1621.3, "median_completion_length": 1586, "min_completion_length": 905, "max_completion_length": 2048, "frac_has_think": 0.84, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.18374867586898375, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.146803617477417, "avg_step_time_sec": 28.452215254306793, "gpu_allocated_gb": 32.365616128, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.156304384, "elapsed_sec": 227.66796398162842, "eta_sec": 5462.825328826904, "timestamp": "2026-02-15T13:37:57.276209"}
{"step": 9, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.699609375, "mean_total_reward": 0.7100390625, "std_total_reward": 0.42867333997472984, "total_tokens": 71640, "mean_completion_length": 1432.8, "median_completion_length": 1396, "min_completion_length": 609, "max_completion_length": 2048, "frac_has_think": 0.84, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.07020994158717414, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 27.41387104988098, "avg_step_time_sec": 28.336843676037258, "gpu_allocated_gb": 32.365641728, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.156304384, "elapsed_sec": 255.0890817642212, "eta_sec": 5412.337142123116, "timestamp": "2026-02-15T13:38:24.129583"}
{"step": 10, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.910205078125, "mean_total_reward": 0.5089794921875, "std_total_reward": 0.497293303529208, "total_tokens": 93205, "mean_completion_length": 1864.1, "median_completion_length": 1996, "min_completion_length": 1247, "max_completion_length": 2048, "frac_has_think": 0.72, "frac_has_boxed": 0.72, "mean_reward_std_per_prompt": 0.4036616764129932, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.18489122390747, "avg_step_time_sec": 28.22164843082428, "gpu_allocated_gb": 32.365712384, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 282.2806565761566, "eta_sec": 5362.113201856613, "timestamp": "2026-02-15T13:38:52.344355"}
{"step": 11, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.46, "mean_correctness_reward": 0.46, "mean_length_penalty": -0.8545703125, "mean_total_reward": 0.37454296875, "std_total_reward": 0.514255006011943, "total_tokens": 87508, "mean_completion_length": 1750.16, "median_completion_length": 2048, "min_completion_length": 872, "max_completion_length": 2048, "frac_has_think": 0.6, "frac_has_boxed": 0.6, "mean_reward_std_per_prompt": 0.4242128681595913, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.866064071655273, "avg_step_time_sec": 28.189322579990733, "gpu_allocated_gb": 32.36563456, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 310.153621673584, "eta_sec": 5327.781967618249, "timestamp": "2026-02-15T13:39:20.307181"}
{"step": 12, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.52, "mean_correctness_reward": 0.52, "mean_length_penalty": -0.834716796875, "mean_total_reward": 0.4365283203125, "std_total_reward": 0.5082227783390627, "total_tokens": 85475, "mean_completion_length": 1709.5, "median_completion_length": 1863, "min_completion_length": 789, "max_completion_length": 2048, "frac_has_think": 0.76, "frac_has_boxed": 0.76, "mean_reward_std_per_prompt": 0.3011023013763675, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 27.979257822036743, "avg_step_time_sec": 28.171817183494568, "gpu_allocated_gb": 32.365643776, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 338.1395933628082, "eta_sec": 5296.301630496979, "timestamp": "2026-02-15T13:39:47.594103"}
{"step": 13, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.854482421875, "mean_total_reward": 0.3945517578125, "std_total_reward": 0.5148340936744479, "total_tokens": 87499, "mean_completion_length": 1749.98, "median_completion_length": 2048, "min_completion_length": 986, "max_completion_length": 2048, "frac_has_think": 0.58, "frac_has_boxed": 0.6, "mean_reward_std_per_prompt": 0.29728753442234634, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 27.254947900772095, "avg_step_time_sec": 28.101288777131302, "gpu_allocated_gb": 32.365641728, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 365.4012043476105, "eta_sec": 5254.941001323554, "timestamp": "2026-02-15T13:40:15.623598"}
{"step": 14, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.52, "mean_correctness_reward": 0.52, "mean_length_penalty": -0.743916015625, "mean_total_reward": 0.4456083984375, "std_total_reward": 0.5082698452673549, "total_tokens": 76177, "mean_completion_length": 1523.54, "median_completion_length": 1471, "min_completion_length": 868, "max_completion_length": 2048, "frac_has_think": 0.92, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.20859747619533897, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 28.080584287643433, "avg_step_time_sec": 28.099809885025024, "gpu_allocated_gb": 32.365654016, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 393.48904609680176, "eta_sec": 5226.5646386146545, "timestamp": "2026-02-15T13:40:42.095920"}
{"step": 15, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.81564453125, "mean_total_reward": 0.458435546875, "std_total_reward": 0.5155170057139112, "total_tokens": 83522, "mean_completion_length": 1670.44, "median_completion_length": 1749, "min_completion_length": 764, "max_completion_length": 2048, "frac_has_think": 0.66, "frac_has_boxed": 0.66, "mean_reward_std_per_prompt": 0.22857220279319945, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 26.976412296295166, "avg_step_time_sec": 28.024916712443034, "gpu_allocated_gb": 32.36562944, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 420.47213554382324, "eta_sec": 5184.609591801961, "timestamp": "2026-02-15T13:41:09.685195"}
{"step": 16, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.38, "mean_correctness_reward": 0.38, "mean_length_penalty": -0.891513671875, "mean_total_reward": 0.2908486328125, "std_total_reward": 0.49938384878652275, "total_tokens": 91291, "mean_completion_length": 1825.82, "median_completion_length": 2048, "min_completion_length": 777, "max_completion_length": 2048, "frac_has_think": 0.48, "frac_has_boxed": 0.48, "mean_reward_std_per_prompt": 0.3379542284163225, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 26.998787879943848, "avg_step_time_sec": 27.960783660411835, "gpu_allocated_gb": 32.365674496, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 447.47687816619873, "eta_sec": 5144.784193515778, "timestamp": "2026-02-15T13:41:37.413301"}
{"step": 17, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.845830078125, "mean_total_reward": 0.4554169921875, "std_total_reward": 0.5123487802484832, "total_tokens": 86613, "mean_completion_length": 1732.26, "median_completion_length": 1910, "min_completion_length": 1000, "max_completion_length": 2048, "frac_has_think": 0.72, "frac_has_boxed": 0.74, "mean_reward_std_per_prompt": 0.2258763407693049, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.151695251464844, "avg_step_time_sec": 27.97201375400319, "gpu_allocated_gb": 32.365631488, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 475.63415265083313, "eta_sec": 5118.878516982583, "timestamp": "2026-02-15T13:42:05.103165"}
{"step": 18, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.36, "mean_correctness_reward": 0.36, "mean_length_penalty": -0.92822265625, "mean_total_reward": 0.267177734375, "std_total_reward": 0.48961274248897413, "total_tokens": 95050, "mean_completion_length": 1901, "median_completion_length": 2048, "min_completion_length": 1197, "max_completion_length": 2048, "frac_has_think": 0.46, "frac_has_boxed": 0.46, "mean_reward_std_per_prompt": 0.3954829570266633, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 27.265264987945557, "avg_step_time_sec": 27.932749933666653, "gpu_allocated_gb": 32.36563456, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 502.90556359291077, "eta_sec": 5083.76048792733, "timestamp": "2026-02-15T13:42:33.282370"}
{"step": 19, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.901689453125, "mean_total_reward": 0.3898310546875, "std_total_reward": 0.5088319739403648, "total_tokens": 92333, "mean_completion_length": 1846.66, "median_completion_length": 2048, "min_completion_length": 1186, "max_completion_length": 2048, "frac_has_think": 0.72, "frac_has_boxed": 0.72, "mean_reward_std_per_prompt": 0.36893913455508714, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 28.69491934776306, "avg_step_time_sec": 27.97286411335594, "gpu_allocated_gb": 32.365650944, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 531.6065967082977, "eta_sec": 5063.088404517425, "timestamp": "2026-02-15T13:43:01.569104"}
{"step": 20, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.8271484375, "mean_total_reward": 0.53728515625, "std_total_reward": 0.4989921900571868, "total_tokens": 84700, "mean_completion_length": 1694, "median_completion_length": 1671, "min_completion_length": 1037, "max_completion_length": 2048, "frac_has_think": 0.7, "frac_has_boxed": 0.7, "mean_reward_std_per_prompt": 0.33285190969681167, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 27.64093279838562, "avg_step_time_sec": 27.95626754760742, "gpu_allocated_gb": 32.365620224, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 559.253707408905, "eta_sec": 5032.128158569336, "timestamp": "2026-02-15T13:43:28.661865"}
{"step": 21, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.726044921875, "mean_total_reward": 0.7273955078125, "std_total_reward": 0.41399229560469536, "total_tokens": 74347, "mean_completion_length": 1486.94, "median_completion_length": 1394, "min_completion_length": 693, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.20601925723245373, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.686234712600708, "avg_step_time_sec": 27.94340884117853, "gpu_allocated_gb": 32.365652992, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.261332992, "elapsed_sec": 586.9459915161133, "eta_sec": 5001.870182570957, "timestamp": "2026-02-15T13:43:55.682633"}
{"step": 22, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.46, "mean_correctness_reward": 0.46, "mean_length_penalty": -0.834091796875, "mean_total_reward": 0.3765908203125, "std_total_reward": 0.5122164083137208, "total_tokens": 85411, "mean_completion_length": 1708.22, "median_completion_length": 1936, "min_completion_length": 897, "max_completion_length": 2048, "frac_has_think": 0.68, "frac_has_boxed": 0.68, "mean_reward_std_per_prompt": 0.2476813877976186, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 27.10726571083069, "avg_step_time_sec": 27.90540233525363, "gpu_allocated_gb": 32.365769728, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.392025088, "elapsed_sec": 614.0594792366028, "eta_sec": 4967.161615675146, "timestamp": "2026-02-15T13:44:23.962369"}
{"step": 23, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.46, "mean_correctness_reward": 0.46, "mean_length_penalty": -0.82306640625, "mean_total_reward": 0.377693359375, "std_total_reward": 0.5170803122055895, "total_tokens": 84282, "mean_completion_length": 1685.64, "median_completion_length": 2048, "min_completion_length": 743, "max_completion_length": 2048, "frac_has_think": 0.6, "frac_has_boxed": 0.6, "mean_reward_std_per_prompt": 0.3453999246780042, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 28.624388456344604, "avg_step_time_sec": 27.93666260138802, "gpu_allocated_gb": 32.365832192, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 642.6904211044312, "eta_sec": 4944.78928044568, "timestamp": "2026-02-15T13:44:52.647570"}
{"step": 24, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.7600390625, "mean_total_reward": 0.58399609375, "std_total_reward": 0.491097466815279, "total_tokens": 77828, "mean_completion_length": 1556.56, "median_completion_length": 1539, "min_completion_length": 658, "max_completion_length": 2048, "frac_has_think": 0.74, "frac_has_boxed": 0.74, "mean_reward_std_per_prompt": 0.25328871966199384, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 27.685590744018555, "avg_step_time_sec": 27.926201273997624, "gpu_allocated_gb": 32.365621248, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 670.3827517032623, "eta_sec": 4915.011424223581, "timestamp": "2026-02-15T13:45:19.356953"}
{"step": 25, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.65783203125, "mean_total_reward": 0.414216796875, "std_total_reward": 0.4976333596955848, "total_tokens": 67362, "mean_completion_length": 1347.24, "median_completion_length": 1338, "min_completion_length": 674, "max_completion_length": 2048, "frac_has_think": 0.92, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.17120651752547505, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 26.748103857040405, "avg_step_time_sec": 27.879077377319337, "gpu_allocated_gb": 32.365627392, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 697.1373484134674, "eta_sec": 4878.838541030884, "timestamp": "2026-02-15T13:45:45.708060"}
{"step": 26, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.75677734375, "mean_total_reward": 0.48432226562499997, "std_total_reward": 0.5181104599406879, "total_tokens": 77494, "mean_completion_length": 1549.88, "median_completion_length": 1718, "min_completion_length": 231, "max_completion_length": 2048, "frac_has_think": 0.64, "frac_has_boxed": 0.64, "mean_reward_std_per_prompt": 0.08714375496435194, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 26.43370270729065, "avg_step_time_sec": 27.823486043856693, "gpu_allocated_gb": 32.365635584, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 723.5782570838928, "eta_sec": 4841.286571631064, "timestamp": "2026-02-15T13:46:12.392035"}
{"step": 27, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.62853515625, "mean_total_reward": 0.677146484375, "std_total_reward": 0.443375187927843, "total_tokens": 64362, "mean_completion_length": 1287.24, "median_completion_length": 1155, "min_completion_length": 794, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.1523665018416067, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 26.482728958129883, "avg_step_time_sec": 27.77382837401496, "gpu_allocated_gb": 32.3656192, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 750.0670425891876, "eta_sec": 4804.872308704588, "timestamp": "2026-02-15T13:46:38.283152"}
{"step": 28, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.8009765625, "mean_total_reward": 0.45990234375, "std_total_reward": 0.5170504721200866, "total_tokens": 82020, "mean_completion_length": 1640.4, "median_completion_length": 2037, "min_completion_length": 622, "max_completion_length": 2048, "frac_has_think": 0.58, "frac_has_boxed": 0.58, "mean_reward_std_per_prompt": 0.30520824671838526, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 25.954038619995117, "avg_step_time_sec": 27.708835882799967, "gpu_allocated_gb": 32.365636608, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 776.0269091129303, "eta_sec": 4765.919771841594, "timestamp": "2026-02-15T13:47:05.386896"}
{"step": 29, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.88, "mean_correctness_reward": 0.88, "mean_length_penalty": -0.6601171875, "mean_total_reward": 0.81398828125, "std_total_reward": 0.33780215608214753, "total_tokens": 67596, "mean_completion_length": 1351.92, "median_completion_length": 1251, "min_completion_length": 762, "max_completion_length": 2048, "frac_has_think": 0.92, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.22869883347365527, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.41083002090454, "avg_step_time_sec": 27.698559818596674, "gpu_allocated_gb": 32.365682688, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 803.4445199966431, "eta_sec": 4736.453728980031, "timestamp": "2026-02-15T13:47:32.039832"}
{"step": 30, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.32, "mean_correctness_reward": 0.32, "mean_length_penalty": -0.785, "mean_total_reward": 0.2415, "std_total_reward": 0.46341482947802315, "total_tokens": 80384, "mean_completion_length": 1607.68, "median_completion_length": 1727, "min_completion_length": 542, "max_completion_length": 2048, "frac_has_think": 0.76, "frac_has_boxed": 0.76, "mean_reward_std_per_prompt": 0.29155704743436855, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 26.253264904022217, "avg_step_time_sec": 27.650383321444192, "gpu_allocated_gb": 32.365612032, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 829.7043569087982, "eta_sec": 4700.565164645513, "timestamp": "2026-02-15T13:47:59.023521"}
{"step": 31, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.68, "mean_correctness_reward": 0.68, "mean_length_penalty": -0.661103515625, "mean_total_reward": 0.6138896484375, "std_total_reward": 0.48847742018449325, "total_tokens": 67697, "mean_completion_length": 1353.94, "median_completion_length": 1248, "min_completion_length": 629, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.24185095247290958, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.201156854629517, "avg_step_time_sec": 27.635892145095333, "gpu_allocated_gb": 32.365669376, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 856.9125039577484, "eta_sec": 4670.465772521111, "timestamp": "2026-02-15T13:48:25.038504"}
{"step": 32, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.686435546875, "mean_total_reward": 0.5313564453125, "std_total_reward": 0.4984179979954036, "total_tokens": 70291, "mean_completion_length": 1405.82, "median_completion_length": 1255, "min_completion_length": 709, "max_completion_length": 2048, "frac_has_think": 0.88, "frac_has_boxed": 0.88, "mean_reward_std_per_prompt": 0.1650265061116295, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.77475118637085, "avg_step_time_sec": 27.577731490135193, "gpu_allocated_gb": 32.365612032, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 882.6935663223267, "eta_sec": 4633.058890342712, "timestamp": "2026-02-15T13:48:51.476815"}
{"step": 33, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.68, "mean_correctness_reward": 0.68, "mean_length_penalty": -0.785009765625, "mean_total_reward": 0.6014990234375, "std_total_reward": 0.4813644266455663, "total_tokens": 80385, "mean_completion_length": 1607.7, "median_completion_length": 1603, "min_completion_length": 871, "max_completion_length": 2048, "frac_has_think": 0.68, "frac_has_boxed": 0.7, "mean_reward_std_per_prompt": 0.23896196041568674, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 26.663636445999146, "avg_step_time_sec": 27.550031640312888, "gpu_allocated_gb": 32.365663232, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 909.3632814884186, "eta_sec": 4600.855283932252, "timestamp": "2026-02-15T13:49:18.439866"}
{"step": 34, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.72236328125, "mean_total_reward": 0.627763671875, "std_total_reward": 0.47660794190926475, "total_tokens": 73970, "mean_completion_length": 1479.4, "median_completion_length": 1483, "min_completion_length": 801, "max_completion_length": 2048, "frac_has_think": 0.74, "frac_has_boxed": 0.74, "mean_reward_std_per_prompt": 0.3305769347924317, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 27.166292190551758, "avg_step_time_sec": 27.538745185908148, "gpu_allocated_gb": 32.365696, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 936.5355987548828, "eta_sec": 4571.431700860752, "timestamp": "2026-02-15T13:49:45.553902"}
{"step": 35, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.6365234375, "mean_total_reward": 0.73634765625, "std_total_reward": 0.4183721007790022, "total_tokens": 65180, "mean_completion_length": 1303.6, "median_completion_length": 1255, "min_completion_length": 669, "max_completion_length": 2048, "frac_has_think": 0.92, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.010074479118523917, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 26.78924560546875, "avg_step_time_sec": 27.517330912181308, "gpu_allocated_gb": 32.365632512, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 963.3316721916199, "eta_sec": 4540.359600509916, "timestamp": "2026-02-15T13:50:11.096914"}
{"step": 36, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.65203125, "mean_total_reward": 0.494796875, "std_total_reward": 0.5060331234362212, "total_tokens": 66768, "mean_completion_length": 1335.36, "median_completion_length": 1186, "min_completion_length": 677, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.1270030627138065, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 25.605207681655884, "avg_step_time_sec": 27.464216378000046, "gpu_allocated_gb": 32.365651968, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 988.9428567886353, "eta_sec": 4504.131485992008, "timestamp": "2026-02-15T13:50:37.568808"}
{"step": 37, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.88, "mean_correctness_reward": 0.88, "mean_length_penalty": -0.67234375, "mean_total_reward": 0.812765625, "std_total_reward": 0.3375614110614437, "total_tokens": 68848, "mean_completion_length": 1376.96, "median_completion_length": 1303, "min_completion_length": 749, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.16719796774256576, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.858192443847656, "avg_step_time_sec": 27.447837352752686, "gpu_allocated_gb": 32.36573184, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1015.8071324825287, "eta_sec": 4473.997488498688, "timestamp": "2026-02-15T13:51:03.976527"}
{"step": 38, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.584375, "mean_total_reward": 0.5415625, "std_total_reward": 0.5088899390909907, "total_tokens": 59840, "mean_completion_length": 1196.8, "median_completion_length": 969, "min_completion_length": 605, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.007716755786266844, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 26.034541130065918, "avg_step_time_sec": 27.410645346892508, "gpu_allocated_gb": 32.365658112, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1041.848069190979, "eta_sec": 4440.524546196586, "timestamp": "2026-02-15T13:51:29.224115"}
{"step": 39, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.448984375, "mean_total_reward": 0.9551015625, "std_total_reward": 0.006801281011756458, "total_tokens": 45976, "mean_completion_length": 919.52, "median_completion_length": 913, "min_completion_length": 667, "max_completion_length": 1342, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.005675743768205407, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 25.026526927947998, "avg_step_time_sec": 27.34951410538111, "gpu_allocated_gb": 32.365620224, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1066.8796858787537, "eta_sec": 4403.271770966359, "timestamp": "2026-02-15T13:51:49.239067"}
{"step": 40, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.46, "mean_correctness_reward": 0.46, "mean_length_penalty": -0.661005859375, "mean_total_reward": 0.3938994140625, "std_total_reward": 0.5098328441195819, "total_tokens": 67687, "mean_completion_length": 1353.74, "median_completion_length": 1243, "min_completion_length": 660, "max_completion_length": 2048, "frac_has_think": 0.9, "frac_has_boxed": 0.9, "mean_reward_std_per_prompt": 0.1663577607981403, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 17.105814695358276, "avg_step_time_sec": 27.09342162013054, "gpu_allocated_gb": 32.3646336, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1083.9915165901184, "eta_sec": 4334.947459220886, "timestamp": "2026-02-15T13:52:12.263583"}
{"step": 41, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.9, "mean_correctness_reward": 0.9, "mean_length_penalty": -0.58794921875, "mean_total_reward": 0.841205078125, "std_total_reward": 0.3144519221335204, "total_tokens": 60206, "mean_completion_length": 1204.12, "median_completion_length": 1138, "min_completion_length": 533, "max_completion_length": 2048, "frac_has_think": 0.92, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.17974002732401836, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.411240577697754, "avg_step_time_sec": 27.076783058119982, "gpu_allocated_gb": 32.365696, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1110.4093205928802, "eta_sec": 4305.208506241077, "timestamp": "2026-02-15T13:52:37.869286"}
{"step": 42, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.45220703125, "mean_total_reward": 0.914779296875, "std_total_reward": 0.19551769456520307, "total_tokens": 46306, "mean_completion_length": 926.12, "median_completion_length": 866, "min_completion_length": 515, "max_completion_length": 1564, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08704286231645328, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 25.842047691345215, "avg_step_time_sec": 27.0473845970063, "gpu_allocated_gb": 32.365651968, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1136.2570931911469, "eta_sec": 4273.486766326995, "timestamp": "2026-02-15T13:52:59.374390"}
{"step": 43, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.606220703125, "mean_total_reward": 0.8793779296875, "std_total_reward": 0.24815612590245303, "total_tokens": 62077, "mean_completion_length": 1241.54, "median_completion_length": 1102, "min_completion_length": 586, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.10014642655392561, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 18.73045325279236, "avg_step_time_sec": 26.85396758900132, "gpu_allocated_gb": 32.364928512, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1154.9943389892578, "eta_sec": 4216.072911473208, "timestamp": "2026-02-15T13:53:22.519712"}
{"step": 44, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.64, "mean_correctness_reward": 0.64, "mean_length_penalty": -0.571171875, "mean_total_reward": 0.5828828125, "std_total_reward": 0.5011887518318509, "total_tokens": 58488, "mean_completion_length": 1169.76, "median_completion_length": 1040, "min_completion_length": 465, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.15977175171805655, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.31675910949707, "avg_step_time_sec": 26.819031032648954, "gpu_allocated_gb": 32.365645824, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1180.3162307739258, "eta_sec": 4183.768841093237, "timestamp": "2026-02-15T13:53:47.587110"}
{"step": 45, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.58005859375, "mean_total_reward": 0.741994140625, "std_total_reward": 0.4211087600733525, "total_tokens": 59398, "mean_completion_length": 1187.96, "median_completion_length": 1102, "min_completion_length": 514, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.004555336199880983, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.55101251602173, "avg_step_time_sec": 26.79085284339057, "gpu_allocated_gb": 32.365638656, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1205.874041557312, "eta_sec": 4152.582190725539, "timestamp": "2026-02-15T13:54:13.290127"}
{"step": 46, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.9, "mean_correctness_reward": 0.9, "mean_length_penalty": -0.5025390625, "mean_total_reward": 0.84974609375, "std_total_reward": 0.316915365503492, "total_tokens": 51460, "mean_completion_length": 1029.2, "median_completion_length": 804, "min_completion_length": 646, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.10695584991983376, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 25.28043031692505, "avg_step_time_sec": 26.758017571076103, "gpu_allocated_gb": 32.3656704, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1231.1616117954254, "eta_sec": 4120.73470594572, "timestamp": "2026-02-15T13:54:37.786529"}
{"step": 47, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.68, "mean_correctness_reward": 0.68, "mean_length_penalty": -0.560126953125, "mean_total_reward": 0.6239873046875, "std_total_reward": 0.47566236379889776, "total_tokens": 57357, "mean_completion_length": 1147.14, "median_completion_length": 1053, "min_completion_length": 510, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.18268367211417044, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 24.408305406570435, "avg_step_time_sec": 26.708023695235557, "gpu_allocated_gb": 32.36563456, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1255.5758559703827, "eta_sec": 4086.32762537104, "timestamp": "2026-02-15T13:55:02.684132"}
{"step": 48, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.545068359375, "mean_total_reward": 0.7454931640625, "std_total_reward": 0.4228817165979108, "total_tokens": 55815, "mean_completion_length": 1116.3, "median_completion_length": 955, "min_completion_length": 406, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.00523046998289256, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.422040700912476, "avg_step_time_sec": 26.681232382853825, "gpu_allocated_gb": 32.3656448, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1281.0038969516754, "eta_sec": 4055.5473221937814, "timestamp": "2026-02-15T13:55:27.989726"}
{"step": 49, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.503388671875, "mean_total_reward": 0.5496611328125, "std_total_reward": 0.5074643194321269, "total_tokens": 51547, "mean_completion_length": 1030.94, "median_completion_length": 880, "min_completion_length": 453, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.003649269196730683, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 24.772444009780884, "avg_step_time_sec": 26.642277518097234, "gpu_allocated_gb": 32.365645824, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1305.7826926708221, "eta_sec": 4022.9839052326824, "timestamp": "2026-02-15T13:55:52.511651"}
{"step": 50, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.462294921875, "mean_total_reward": 0.9337705078125, "std_total_reward": 0.1410942700787292, "total_tokens": 47339, "mean_completion_length": 946.78, "median_completion_length": 892, "min_completion_length": 465, "max_completion_length": 1951, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06619344968350706, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 24.51799488067627, "avg_step_time_sec": 26.599791865348816, "gpu_allocated_gb": 32.365627392, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1330.306846857071, "eta_sec": 3989.9687798023224, "timestamp": "2026-02-15T13:56:16.485416"}
{"step": 51, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.4609765625, "mean_total_reward": 0.69390234375, "std_total_reward": 0.44366890541207316, "total_tokens": 47204, "mean_completion_length": 944.08, "median_completion_length": 784, "min_completion_length": 574, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.09812787285734524, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.376965761184692, "avg_step_time_sec": 26.536599196639717, "gpu_allocated_gb": 32.365487616, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1354.2336385250092, "eta_sec": 3953.9532802993176, "timestamp": "2026-02-15T13:56:40.471636"}
{"step": 52, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.56404296875, "mean_total_reward": 0.563595703125, "std_total_reward": 0.504146462053398, "total_tokens": 57758, "mean_completion_length": 1155.16, "median_completion_length": 953, "min_completion_length": 548, "max_completion_length": 2048, "frac_has_think": 0.84, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.21989579481530014, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.888306379318237, "avg_step_time_sec": 26.485670488614304, "gpu_allocated_gb": 32.365621248, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1378.1284668445587, "eta_sec": 3919.879232314917, "timestamp": "2026-02-15T13:57:05.370957"}
{"step": 53, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.643916015625, "mean_total_reward": 0.6356083984374999, "std_total_reward": 0.47591476133734756, "total_tokens": 65937, "mean_completion_length": 1318.74, "median_completion_length": 1294, "min_completion_length": 669, "max_completion_length": 2048, "frac_has_think": 0.78, "frac_has_boxed": 0.78, "mean_reward_std_per_prompt": 0.22425404252248796, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.029269218444824, "avg_step_time_sec": 26.458191219365823, "gpu_allocated_gb": 32.365645824, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1403.1648936271667, "eta_sec": 3889.354109246776, "timestamp": "2026-02-15T13:57:31.796449"}
{"step": 54, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.646787109375, "mean_total_reward": 0.5953212890625, "std_total_reward": 0.4992953589736158, "total_tokens": 66231, "mean_completion_length": 1324.62, "median_completion_length": 1032, "min_completion_length": 520, "max_completion_length": 2048, "frac_has_think": 0.7, "frac_has_boxed": 0.7, "mean_reward_std_per_prompt": 0.14382758351992056, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.48470711708069, "avg_step_time_sec": 26.45868225450869, "gpu_allocated_gb": 32.365657088, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1429.656486272812, "eta_sec": 3862.9676091582687, "timestamp": "2026-02-15T13:57:57.692459"}
{"step": 55, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.5294140625, "mean_total_reward": 0.50705859375, "std_total_reward": 0.5095935668233836, "total_tokens": 54212, "mean_completion_length": 1084.24, "median_completion_length": 771, "min_completion_length": 404, "max_completion_length": 2048, "frac_has_think": 0.76, "frac_has_boxed": 0.76, "mean_reward_std_per_prompt": 0.08477019168763876, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 25.827713012695312, "avg_step_time_sec": 26.447210086475717, "gpu_allocated_gb": 32.365641728, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1455.4907505512238, "eta_sec": 3834.845462538979, "timestamp": "2026-02-15T13:58:22.522126"}
{"step": 56, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.605634765625, "mean_total_reward": 0.6994365234375, "std_total_reward": 0.4496682896554151, "total_tokens": 62017, "mean_completion_length": 1240.34, "median_completion_length": 828, "min_completion_length": 426, "max_completion_length": 2048, "frac_has_think": 0.78, "frac_has_boxed": 0.78, "mean_reward_std_per_prompt": 0.08307100220302832, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 24.825702905654907, "avg_step_time_sec": 26.418254601103918, "gpu_allocated_gb": 32.365648896, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1480.3238170146942, "eta_sec": 3804.228662558964, "timestamp": "2026-02-15T13:58:48.562837"}
{"step": 57, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.69802734375, "mean_total_reward": 0.590197265625, "std_total_reward": 0.49331051145467186, "total_tokens": 71478, "mean_completion_length": 1429.56, "median_completion_length": 1661, "min_completion_length": 603, "max_completion_length": 2048, "frac_has_think": 0.68, "frac_has_boxed": 0.68, "mean_reward_std_per_prompt": 0.22260146816788284, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 26.047309637069702, "avg_step_time_sec": 26.411746794717352, "gpu_allocated_gb": 32.365647872, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1506.379275560379, "eta_sec": 3776.8797916445815, "timestamp": "2026-02-15T13:59:14.902306"}
{"step": 58, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.52462890625, "mean_total_reward": 0.787537109375, "std_total_reward": 0.38763411416782556, "total_tokens": 53722, "mean_completion_length": 1074.44, "median_completion_length": 1009, "min_completion_length": 464, "max_completion_length": 2048, "frac_has_think": 0.9, "frac_has_boxed": 0.9, "mean_reward_std_per_prompt": 0.08728483562587454, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 26.781111240386963, "avg_step_time_sec": 26.4181151472289, "gpu_allocated_gb": 32.36573696, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1533.1668140888214, "eta_sec": 3751.372350906504, "timestamp": "2026-02-15T13:59:39.994505"}
{"step": 59, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.38, "mean_correctness_reward": 0.38, "mean_length_penalty": -0.497646484375, "mean_total_reward": 0.3302353515625, "std_total_reward": 0.4846097055412222, "total_tokens": 50959, "mean_completion_length": 1019.18, "median_completion_length": 1067, "min_completion_length": 509, "max_completion_length": 1697, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06470259504549, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 24.542096376419067, "avg_step_time_sec": 26.386318218910088, "gpu_allocated_gb": 32.365630464, "gpu_reserved_gb": 79.956017152, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1557.7156691551208, "eta_sec": 3720.4708688663222, "timestamp": "2026-02-15T14:00:02.594798"}
{"step": 60, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.433154296875, "mean_total_reward": 0.9566845703125, "std_total_reward": 0.010553773107669437, "total_tokens": 44355, "mean_completion_length": 887.1, "median_completion_length": 836, "min_completion_length": 516, "max_completion_length": 1610, "frac_has_think": 1.0, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.006784300513421201, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 21.37665820121765, "avg_step_time_sec": 26.30282388528188, "gpu_allocated_gb": 32.365159936, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1579.098572254181, "eta_sec": 3682.3953439394636, "timestamp": "2026-02-15T14:00:22.342786"}
{"step": 61, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.593349609375, "mean_total_reward": 0.7806650390625, "std_total_reward": 0.3802993811633672, "total_tokens": 60759, "mean_completion_length": 1215.18, "median_completion_length": 1285, "min_completion_length": 451, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.24697190148283252, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 19.14186692237854, "avg_step_time_sec": 26.185431148185106, "gpu_allocated_gb": 32.36500736, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1598.2467584609985, "eta_sec": 3639.7749295977296, "timestamp": "2026-02-15T14:00:45.670304"}
{"step": 62, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.45033203125, "mean_total_reward": 0.754966796875, "std_total_reward": 0.4178994992554625, "total_tokens": 46114, "mean_completion_length": 922.28, "median_completion_length": 852, "min_completion_length": 431, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.005305553070251189, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 25.27052068710327, "avg_step_time_sec": 26.170674527845076, "gpu_allocated_gb": 32.365659136, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1623.5229189395905, "eta_sec": 3611.5530848426206, "timestamp": "2026-02-15T14:01:10.182683"}
{"step": 63, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.460810546875, "mean_total_reward": 0.9139189453125, "std_total_reward": 0.19792017542596518, "total_tokens": 47187, "mean_completion_length": 943.74, "median_completion_length": 957, "min_completion_length": 565, "max_completion_length": 1313, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12372370126841462, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 24.527483463287354, "avg_step_time_sec": 26.14459212999495, "gpu_allocated_gb": 32.365652992, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1648.0562317371368, "eta_sec": 3581.809121809308, "timestamp": "2026-02-15T14:01:29.821772"}
{"step": 64, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.9, "mean_correctness_reward": 0.9, "mean_length_penalty": -0.529814453125, "mean_total_reward": 0.8470185546875, "std_total_reward": 0.31466886136146993, "total_tokens": 54253, "mean_completion_length": 1085.06, "median_completion_length": 966, "min_completion_length": 466, "max_completion_length": 2048, "frac_has_think": 0.92, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.16382398567360354, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 16.489157915115356, "avg_step_time_sec": 25.99372597038746, "gpu_allocated_gb": 32.364615168, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1664.5517048835754, "eta_sec": 3535.1467319726944, "timestamp": "2026-02-15T14:01:51.486214"}
{"step": 65, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.464384765625, "mean_total_reward": 0.7335615234375, "std_total_reward": 0.44017903723456364, "total_tokens": 47553, "mean_completion_length": 951.06, "median_completion_length": 770, "min_completion_length": 395, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.061526675177579676, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 24.788190841674805, "avg_step_time_sec": 25.97517927609957, "gpu_allocated_gb": 32.36566016, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1689.3459031581879, "eta_sec": 3506.649202273442, "timestamp": "2026-02-15T14:02:16.296500"}
{"step": 66, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.471396484375, "mean_total_reward": 0.5328603515625, "std_total_reward": 0.49783458064380337, "total_tokens": 48271, "mean_completion_length": 965.42, "median_completion_length": 977, "min_completion_length": 309, "max_completion_length": 1684, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06455330712767507, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 24.7770574092865, "avg_step_time_sec": 25.957025914481193, "gpu_allocated_gb": 32.36565504, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1714.1298410892487, "eta_sec": 3478.2414725404797, "timestamp": "2026-02-15T14:02:38.237318"}
{"step": 67, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.465302734375, "mean_total_reward": 0.7534697265625, "std_total_reward": 0.4117903582105596, "total_tokens": 47647, "mean_completion_length": 952.94, "median_completion_length": 859, "min_completion_length": 565, "max_completion_length": 1588, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.004172379542565112, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 20.3342182636261, "avg_step_time_sec": 25.873103412229625, "gpu_allocated_gb": 32.36512768, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1734.470342874527, "eta_sec": 3441.1227538265402, "timestamp": "2026-02-15T14:02:57.950189"}
{"step": 68, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.471962890625, "mean_total_reward": 0.5528037109375, "std_total_reward": 0.5057975974367535, "total_tokens": 48329, "mean_completion_length": 966.58, "median_completion_length": 713, "min_completion_length": 532, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.0040202720449702915, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 19.398823499679565, "avg_step_time_sec": 25.777893413515653, "gpu_allocated_gb": 32.365005312, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1753.8753833770752, "eta_sec": 3402.6819305840663, "timestamp": "2026-02-15T14:03:20.871278"}
{"step": 69, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.2, "mean_correctness_reward": 0.2, "mean_length_penalty": -0.581708984375, "mean_total_reward": 0.1418291015625, "std_total_reward": 0.41216407097437335, "total_tokens": 59567, "mean_completion_length": 1191.34, "median_completion_length": 864, "min_completion_length": 434, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.003792482477454189, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.8, "step_time_sec": 24.81372880935669, "avg_step_time_sec": 25.763920013455376, "gpu_allocated_gb": 32.365641728, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1778.6963403224945, "eta_sec": 3375.0735217626543, "timestamp": "2026-02-15T14:03:46.152838"}
{"step": 70, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.419267578125, "mean_total_reward": 0.5380732421875, "std_total_reward": 0.5002679405226728, "total_tokens": 42933, "mean_completion_length": 858.66, "median_completion_length": 741, "min_completion_length": 420, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.06862222664707084, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 25.287508487701416, "avg_step_time_sec": 25.757114134516033, "gpu_allocated_gb": 32.36564992, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1803.9907562732697, "eta_sec": 3348.4248374870845, "timestamp": "2026-02-15T14:04:09.897958"}
{"step": 71, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.370927734375, "mean_total_reward": 0.7629072265625, "std_total_reward": 0.4074831951643113, "total_tokens": 37983, "mean_completion_length": 759.66, "median_completion_length": 687, "min_completion_length": 546, "max_completion_length": 1731, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.004908698665582212, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.552327632904053, "avg_step_time_sec": 25.726060803507416, "gpu_allocated_gb": 32.365621248, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1827.5491228103638, "eta_sec": 3318.661843652457, "timestamp": "2026-02-15T14:04:30.907542"}
{"step": 72, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.49646484375, "mean_total_reward": 0.550353515625, "std_total_reward": 0.505879416527308, "total_tokens": 50838, "mean_completion_length": 1016.76, "median_completion_length": 806, "min_completion_length": 481, "max_completion_length": 2048, "frac_has_think": 0.8, "frac_has_boxed": 0.8, "mean_reward_std_per_prompt": 0.003644066377881783, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 19.912566900253296, "avg_step_time_sec": 25.645317832628887, "gpu_allocated_gb": 32.3652224, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1847.4681622982025, "eta_sec": 3282.6006825764975, "timestamp": "2026-02-15T14:04:54.674277"}
{"step": 73, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.431640625, "mean_total_reward": 0.3568359375, "std_total_reward": 0.5001525656633851, "total_tokens": 44200, "mean_completion_length": 884, "median_completion_length": 771, "min_completion_length": 370, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.008057015977102218, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 25.02824592590332, "avg_step_time_sec": 25.636864792810726, "gpu_allocated_gb": 32.36562944, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1872.5027770996094, "eta_sec": 3255.8818286869623, "timestamp": "2026-02-15T14:05:18.628373"}
{"step": 74, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.40546875, "mean_total_reward": 0.799453125, "std_total_reward": 0.3779395424516541, "total_tokens": 41520, "mean_completion_length": 830.4, "median_completion_length": 651, "min_completion_length": 278, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.15728398619152756, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.960726976394653, "avg_step_time_sec": 25.61421428177808, "gpu_allocated_gb": 32.365658112, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1896.4693973064423, "eta_sec": 3227.390999504038, "timestamp": "2026-02-15T14:05:42.276126"}
{"step": 75, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.397861328125, "mean_total_reward": 0.7402138671875, "std_total_reward": 0.43811923362975047, "total_tokens": 40741, "mean_completion_length": 814.82, "median_completion_length": 736, "min_completion_length": 150, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.06663639155568465, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.549935579299927, "avg_step_time_sec": 25.586690565745037, "gpu_allocated_gb": 32.365628416, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1920.0257444381714, "eta_sec": 3198.3363207181296, "timestamp": "2026-02-15T14:06:06.425085"}
{"step": 76, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.86, "mean_correctness_reward": 0.86, "mean_length_penalty": -0.37064453125, "mean_total_reward": 0.822935546875, "std_total_reward": 0.3638872013752435, "total_tokens": 37954, "mean_completion_length": 759.08, "median_completion_length": 611, "min_completion_length": 257, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.09552687314743312, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 24.22849440574646, "avg_step_time_sec": 25.568819563639792, "gpu_allocated_gb": 32.365657088, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1944.260358095169, "eta_sec": 3170.5336258913344, "timestamp": "2026-02-15T14:06:29.788278"}
{"step": 77, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.414111328125, "mean_total_reward": 0.5585888671875, "std_total_reward": 0.5040487954960362, "total_tokens": 42405, "mean_completion_length": 848.1, "median_completion_length": 662, "min_completion_length": 407, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.006929507756177072, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 23.267491817474365, "avg_step_time_sec": 25.538932190312966, "gpu_allocated_gb": 32.36562432, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1967.5338945388794, "eta_sec": 3141.288659408495, "timestamp": "2026-02-15T14:06:53.400670"}
{"step": 78, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.403447265625, "mean_total_reward": 0.7196552734375, "std_total_reward": 0.42688203017895343, "total_tokens": 41313, "mean_completion_length": 826.26, "median_completion_length": 795, "min_completion_length": 454, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.1250023961662828, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.640998363494873, "avg_step_time_sec": 25.51459970535376, "gpu_allocated_gb": 32.365625344, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 1991.180776834488, "eta_sec": 3112.7811640531586, "timestamp": "2026-02-15T14:07:16.584115"}
{"step": 79, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.3423046875, "mean_total_reward": 0.36576953125, "std_total_reward": 0.494841992269176, "total_tokens": 35052, "mean_completion_length": 701.04, "median_completion_length": 685, "min_completion_length": 462, "max_completion_length": 1760, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.005515976318995195, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 23.640249252319336, "avg_step_time_sec": 25.49087375025206, "gpu_allocated_gb": 32.365647872, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2014.826523065567, "eta_sec": 3084.3957237804993, "timestamp": "2026-02-15T14:07:37.918028"}
{"step": 80, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.40072265625, "mean_total_reward": 0.619927734375, "std_total_reward": 0.4673521450277142, "total_tokens": 41034, "mean_completion_length": 820.68, "median_completion_length": 834, "min_completion_length": 431, "max_completion_length": 1296, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.19311642649302288, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 19.735042810440063, "avg_step_time_sec": 25.41892586350441, "gpu_allocated_gb": 32.365248512, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2034.5676884651184, "eta_sec": 3050.271103620529, "timestamp": "2026-02-15T14:07:55.907943"}
{"step": 81, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.268916015625, "mean_total_reward": 0.7531083984375, "std_total_reward": 0.4198250290995992, "total_tokens": 27537, "mean_completion_length": 550.74, "median_completion_length": 513, "min_completion_length": 311, "max_completion_length": 886, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06204037155766738, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 16.421472549438477, "avg_step_time_sec": 25.307846192960387, "gpu_allocated_gb": 32.3646848, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2050.995434522629, "eta_sec": 3011.633696962286, "timestamp": "2026-02-15T14:08:08.738599"}
{"step": 82, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.271904296875, "mean_total_reward": 0.5128095703125, "std_total_reward": 0.4979872692014835, "total_tokens": 27843, "mean_completion_length": 556.86, "median_completion_length": 539, "min_completion_length": 349, "max_completion_length": 1133, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0937435995933938, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 10.904476881027222, "avg_step_time_sec": 25.132195347692907, "gpu_allocated_gb": 32.364012032, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2061.908364534378, "eta_sec": 2965.599051027763, "timestamp": "2026-02-15T14:08:20.985599"}
{"step": 83, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.330029296875, "mean_total_reward": 0.7669970703125, "std_total_reward": 0.3950110694988033, "total_tokens": 33795, "mean_completion_length": 675.9, "median_completion_length": 466, "min_completion_length": 288, "max_completion_length": 1695, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.00355788525264212, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 13.15932583808899, "avg_step_time_sec": 24.987943907818163, "gpu_allocated_gb": 32.364350976, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2075.0748670101166, "eta_sec": 2923.589437214725, "timestamp": "2026-02-15T14:08:38.734314"}
{"step": 84, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.460087890625, "mean_total_reward": 0.5539912109375, "std_total_reward": 0.5045790867823927, "total_tokens": 47113, "mean_completion_length": 942.26, "median_completion_length": 548, "min_completion_length": 308, "max_completion_length": 2048, "frac_has_think": 0.9, "frac_has_boxed": 0.9, "mean_reward_std_per_prompt": 0.20608526103289992, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 20.087766408920288, "avg_step_time_sec": 24.92960846140271, "gpu_allocated_gb": 32.365168128, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2095.168345451355, "eta_sec": 2891.8345815227144, "timestamp": "2026-02-15T14:09:01.497609"}
{"step": 85, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.19740234375, "mean_total_reward": 0.9802597656250001, "std_total_reward": 0.005731224254687609, "total_tokens": 20214, "mean_completion_length": 404.28, "median_completion_length": 409, "min_completion_length": 213, "max_completion_length": 726, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.002330184779494538, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.943610668182373, "avg_step_time_sec": 24.918008487364826, "gpu_allocated_gb": 32.365607936, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2119.117956161499, "eta_sec": 2865.570976046955, "timestamp": "2026-02-15T14:09:15.324849"}
{"step": 86, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.305947265625, "mean_total_reward": 0.5694052734375, "std_total_reward": 0.4943225059263467, "total_tokens": 31329, "mean_completion_length": 626.58, "median_completion_length": 640, "min_completion_length": 209, "max_completion_length": 1219, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.005335676651108656, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 8.730660438537598, "avg_step_time_sec": 24.72978351005288, "gpu_allocated_gb": 32.363763712, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2127.855204343796, "eta_sec": 2819.1953201460283, "timestamp": "2026-02-15T14:09:27.783169"}
{"step": 87, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.240029296875, "mean_total_reward": 0.9759970703125, "std_total_reward": 0.00845783800325296, "total_tokens": 24579, "mean_completion_length": 491.58, "median_completion_length": 471, "min_completion_length": 249, "max_completion_length": 894, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.003778472680266659, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 14.324642181396484, "avg_step_time_sec": 24.61018418443614, "gpu_allocated_gb": 32.364460544, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2142.1867852211, "eta_sec": 2780.9508128412836, "timestamp": "2026-02-15T14:09:40.166716"}
{"step": 88, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.362890625, "mean_total_reward": 0.6237109375, "std_total_reward": 0.4851621042288451, "total_tokens": 37160, "mean_completion_length": 743.2, "median_completion_length": 648, "min_completion_length": 250, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0936474166292172, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 11.223520755767822, "avg_step_time_sec": 24.458063009110365, "gpu_allocated_gb": 32.364040192, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2153.4159948825836, "eta_sec": 2739.303057020361, "timestamp": "2026-02-15T14:09:58.962041"}
{"step": 89, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.92, "mean_correctness_reward": 0.92, "mean_length_penalty": -0.42169921875, "mean_total_reward": 0.877830078125, "std_total_reward": 0.2820149708893033, "total_tokens": 43182, "mean_completion_length": 863.64, "median_completion_length": 653, "min_completion_length": 362, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.203793908985887, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.413082122802734, "avg_step_time_sec": 24.446321650837245, "gpu_allocated_gb": 32.365671424, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2176.8354728221893, "eta_sec": 2713.541703242934, "timestamp": "2026-02-15T14:10:22.871733"}
{"step": 90, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.33873046875, "mean_total_reward": 0.906126953125, "std_total_reward": 0.248587771712948, "total_tokens": 34686, "mean_completion_length": 693.72, "median_completion_length": 468, "min_completion_length": 98, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.14305755427490885, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.827696561813354, "avg_step_time_sec": 24.43944803873698, "gpu_allocated_gb": 32.365637632, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2200.6696820259094, "eta_sec": 2688.339284261068, "timestamp": "2026-02-15T14:10:46.136018"}
{"step": 91, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.34294921875, "mean_total_reward": 0.685705078125, "std_total_reward": 0.452577750714103, "total_tokens": 35118, "mean_completion_length": 702.36, "median_completion_length": 457, "min_completion_length": 167, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.15414248631405383, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 24.004480838775635, "avg_step_time_sec": 24.434668179396745, "gpu_allocated_gb": 32.365791232, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2224.680882692337, "eta_sec": 2663.378831554245, "timestamp": "2026-02-15T14:11:10.624758"}
{"step": 92, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.470078125, "mean_total_reward": 0.6929921875, "std_total_reward": 0.4594135214544963, "total_tokens": 48136, "mean_completion_length": 962.72, "median_completion_length": 806, "min_completion_length": 336, "max_completion_length": 2048, "frac_has_think": 0.88, "frac_has_boxed": 0.88, "mean_reward_std_per_prompt": 0.16117528626791777, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.74741768836975, "avg_step_time_sec": 24.427198065363843, "gpu_allocated_gb": 32.365626368, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2248.434865951538, "eta_sec": 2638.137391059295, "timestamp": "2026-02-15T14:11:35.111224"}
{"step": 93, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.32662109375, "mean_total_reward": 0.447337890625, "std_total_reward": 0.49711639041941347, "total_tokens": 33446, "mean_completion_length": 668.92, "median_completion_length": 510, "min_completion_length": 306, "max_completion_length": 1606, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.10223714679381907, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 25.0278263092041, "avg_step_time_sec": 24.433656433577177, "gpu_allocated_gb": 32.365775872, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2273.469218015671, "eta_sec": 2614.401238392758, "timestamp": "2026-02-15T14:11:56.157450"}
{"step": 94, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.214296875, "mean_total_reward": 0.9785703125, "std_total_reward": 0.006104161215856768, "total_tokens": 21944, "mean_completion_length": 438.88, "median_completion_length": 422, "min_completion_length": 216, "max_completion_length": 730, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0024797341201173344, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 19.411589860916138, "avg_step_time_sec": 24.380230193442486, "gpu_allocated_gb": 32.36520448, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2292.8877704143524, "eta_sec": 2584.3044005049037, "timestamp": "2026-02-15T14:12:09.223221"}
{"step": 95, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.254814453125, "mean_total_reward": 0.7545185546875, "std_total_reward": 0.4214974993578401, "total_tokens": 26093, "mean_completion_length": 521.86, "median_completion_length": 541, "min_completion_length": 197, "max_completion_length": 961, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06359443435692916, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.902240514755249, "avg_step_time_sec": 24.21730398629841, "gpu_allocated_gb": 32.363778048, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2301.7959525585175, "eta_sec": 2542.8169185613333, "timestamp": "2026-02-15T14:12:20.384953"}
{"step": 96, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.230166015625, "mean_total_reward": 0.9569833984375, "std_total_reward": 0.14082860033595637, "total_tokens": 23569, "mean_completion_length": 471.38, "median_completion_length": 446, "min_completion_length": 218, "max_completion_length": 951, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.061759039705696926, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 12.160884380340576, "avg_step_time_sec": 24.091716282069683, "gpu_allocated_gb": 32.364132352, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2313.9627254009247, "eta_sec": 2505.538493335247, "timestamp": "2026-02-15T14:12:31.752414"}
{"step": 97, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.219755859375, "mean_total_reward": 0.3780244140625, "std_total_reward": 0.4988264001836508, "total_tokens": 22503, "mean_completion_length": 450.06, "median_completion_length": 318, "min_completion_length": 151, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.006971540932823718, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 11.235397577285767, "avg_step_time_sec": 23.959176913979128, "gpu_allocated_gb": 32.3641088, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2325.2040605545044, "eta_sec": 2467.79522213985, "timestamp": "2026-02-15T14:12:49.515874"}
{"step": 98, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.160380859375, "mean_total_reward": 0.7439619140625, "std_total_reward": 0.4278001902055518, "total_tokens": 16423, "mean_completion_length": 328.46, "median_completion_length": 329, "min_completion_length": 164, "max_completion_length": 621, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12134056330634102, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.08588457107544, "avg_step_time_sec": 23.940061685990315, "gpu_allocated_gb": 32.365626368, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2347.2961280345917, "eta_sec": 2441.8862919710123, "timestamp": "2026-02-15T14:13:02.550708"}
{"step": 99, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.145419921875, "mean_total_reward": 0.9454580078125, "std_total_reward": 0.19649772813808628, "total_tokens": 14891, "mean_completion_length": 297.82, "median_completion_length": 283, "min_completion_length": 161, "max_completion_length": 691, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12059006319414524, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.507134437561035, "avg_step_time_sec": 23.774072521864767, "gpu_allocated_gb": 32.36364544, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2354.809501886368, "eta_sec": 2401.1813247083414, "timestamp": "2026-02-15T14:13:10.295275"}
{"step": 100, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.22720703125, "mean_total_reward": 0.957279296875, "std_total_reward": 0.15168521839464502, "total_tokens": 23266, "mean_completion_length": 465.32, "median_completion_length": 365, "min_completion_length": 215, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.06546076615030788, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.875882148742676, "avg_step_time_sec": 23.615090618133546, "gpu_allocated_gb": 32.363716608, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2362.6913723945618, "eta_sec": 2361.5090618133545, "timestamp": "2026-02-15T14:13:27.528641"}
{"step": 101, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.219794921875, "mean_total_reward": 0.7580205078125, "std_total_reward": 0.42081667584843496, "total_tokens": 22507, "mean_completion_length": 450.14, "median_completion_length": 322, "min_completion_length": 131, "max_completion_length": 1947, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.27943112270070364, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.60590386390686, "avg_step_time_sec": 23.605098670071893, "gpu_allocated_gb": 32.365668352, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2385.8314878940582, "eta_sec": 2336.9047683371173, "timestamp": "2026-02-15T14:13:49.162411"}
{"step": 102, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.165634765625, "mean_total_reward": 0.7834365234375, "std_total_reward": 0.405023845664087, "total_tokens": 16961, "mean_completion_length": 339.22, "median_completion_length": 325, "min_completion_length": 121, "max_completion_length": 964, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.003493832460512782, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 20.80250358581543, "avg_step_time_sec": 23.577622247677223, "gpu_allocated_gb": 32.36551936, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2406.6406157016754, "eta_sec": 2310.6069802723678, "timestamp": "2026-02-15T14:14:04.048095"}
{"step": 103, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.34, "mean_correctness_reward": 0.34, "mean_length_penalty": -0.37046875, "mean_total_reward": 0.302953125, "std_total_reward": 0.48104862807985915, "total_tokens": 37936, "mean_completion_length": 758.72, "median_completion_length": 715, "min_completion_length": 216, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.1047520401764081, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 11.255768299102783, "avg_step_time_sec": 23.457992597691064, "gpu_allocated_gb": 32.364217856, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2417.9042649269104, "eta_sec": 2275.425281976033, "timestamp": "2026-02-15T14:14:23.481915"}
{"step": 104, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.170654296875, "mean_total_reward": 0.5829345703125, "std_total_reward": 0.4937836661389842, "total_tokens": 17475, "mean_completion_length": 349.5, "median_completion_length": 340, "min_completion_length": 123, "max_completion_length": 611, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.250478161396236, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.360639572143555, "avg_step_time_sec": 23.457056510906952, "gpu_allocated_gb": 32.365656064, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2441.2711250782013, "eta_sec": 2251.8774250470674, "timestamp": "2026-02-15T14:14:36.573878"}
{"step": 105, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.92, "mean_correctness_reward": 0.92, "mean_length_penalty": -0.15814453125, "mean_total_reward": 0.904185546875, "std_total_reward": 0.2740880535465826, "total_tokens": 16194, "mean_completion_length": 323.88, "median_completion_length": 264, "min_completion_length": 131, "max_completion_length": 821, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1519259742820898, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.45745325088501, "avg_step_time_sec": 23.304679337001982, "gpu_allocated_gb": 32.363613696, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2448.7350976467133, "eta_sec": 2213.9445370151884, "timestamp": "2026-02-15T14:14:45.165331"}
{"step": 106, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.324267578125, "mean_total_reward": 0.5675732421874999, "std_total_reward": 0.5086907688546247, "total_tokens": 33205, "mean_completion_length": 664.1, "median_completion_length": 437, "min_completion_length": 174, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.0037366027116939324, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 9.309100866317749, "avg_step_time_sec": 23.172645577844584, "gpu_allocated_gb": 32.363914752, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2458.0502829551697, "eta_sec": 2178.228684317391, "timestamp": "2026-02-15T14:15:03.973218"}
{"step": 107, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.17650390625, "mean_total_reward": 0.522349609375, "std_total_reward": 0.500116473273271, "total_tokens": 18074, "mean_completion_length": 361.48, "median_completion_length": 306, "min_completion_length": 155, "max_completion_length": 695, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09356799227459793, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 24.047087907791138, "avg_step_time_sec": 23.180817936068383, "gpu_allocated_gb": 32.365712384, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2482.1032452583313, "eta_sec": 2155.8160680543597, "timestamp": "2026-02-15T14:15:17.966825"}
{"step": 108, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.68, "mean_correctness_reward": 0.68, "mean_length_penalty": -0.216171875, "mean_total_reward": 0.6583828125, "std_total_reward": 0.482904424163402, "total_tokens": 22136, "mean_completion_length": 442.72, "median_completion_length": 232, "min_completion_length": 115, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.2350256363170997, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.499009609222412, "avg_step_time_sec": 23.044875266375364, "gpu_allocated_gb": 32.363761664, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2490.608915567398, "eta_sec": 2120.1285245065333, "timestamp": "2026-02-15T14:15:34.689918"}
{"step": 109, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.132607421875, "mean_total_reward": 0.9667392578125, "std_total_reward": 0.14086314900399635, "total_tokens": 13579, "mean_completion_length": 271.58, "median_completion_length": 267, "min_completion_length": 106, "max_completion_length": 502, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0622912692403295, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 21.72009801864624, "avg_step_time_sec": 23.03272134667143, "gpu_allocated_gb": 32.3656192, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2512.334783554077, "eta_sec": 2095.9776425471, "timestamp": "2026-02-15T14:15:46.722831"}
{"step": 110, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.1206640625, "mean_total_reward": 0.94793359375, "std_total_reward": 0.19646894073598423, "total_tokens": 12356, "mean_completion_length": 247.12, "median_completion_length": 229, "min_completion_length": 109, "max_completion_length": 436, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0815356164422617, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.179744720458984, "avg_step_time_sec": 22.879512468251313, "gpu_allocated_gb": 32.363465216, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2518.5212309360504, "eta_sec": 2059.156122142618, "timestamp": "2026-02-15T14:15:52.438698"}
{"step": 111, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.151416015625, "mean_total_reward": 0.3848583984375, "std_total_reward": 0.49286118563826264, "total_tokens": 15505, "mean_completion_length": 310.1, "median_completion_length": 238, "min_completion_length": 113, "max_completion_length": 1531, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12178291344955308, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 5.556310176849365, "avg_step_time_sec": 22.72344758274319, "gpu_allocated_gb": 32.363390464, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2524.0835180282593, "eta_sec": 2022.386834864144, "timestamp": "2026-02-15T14:16:04.464231"}
{"step": 112, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.090458984375, "mean_total_reward": 0.7909541015625, "std_total_reward": 0.40156536695711165, "total_tokens": 9263, "mean_completion_length": 185.26, "median_completion_length": 174, "min_completion_length": 72, "max_completion_length": 359, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0018786170298086377, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 16.475377798080444, "avg_step_time_sec": 22.667661245380128, "gpu_allocated_gb": 32.364924416, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2540.565037250519, "eta_sec": 1994.7541895934512, "timestamp": "2026-02-15T14:16:13.837502"}
{"step": 113, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.10193359375, "mean_total_reward": 0.829806640625, "std_total_reward": 0.36765357735599075, "total_tokens": 10438, "mean_completion_length": 208.76, "median_completion_length": 178, "min_completion_length": 80, "max_completion_length": 350, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.30360925862142324, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.736472129821777, "avg_step_time_sec": 22.508978155861914, "gpu_allocated_gb": 32.363300352, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2545.307206392288, "eta_sec": 1958.2810995599866, "timestamp": "2026-02-15T14:16:18.574439"}
{"step": 114, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.096328125, "mean_total_reward": 0.7703671875, "std_total_reward": 0.4155843984680621, "total_tokens": 9864, "mean_completion_length": 197.28, "median_completion_length": 181, "min_completion_length": 88, "max_completion_length": 515, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2781836065834169, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.6140546798706055, "avg_step_time_sec": 22.35200514291462, "gpu_allocated_gb": 32.363258368, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2549.9268622398376, "eta_sec": 1922.2724422906574, "timestamp": "2026-02-15T14:16:24.074377"}
{"step": 115, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.083154296875, "mean_total_reward": 0.9516845703125, "std_total_reward": 0.1973842588938268, "total_tokens": 8515, "mean_completion_length": 170.3, "median_completion_length": 159, "min_completion_length": 71, "max_completion_length": 327, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08139703935073929, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.038001775741577, "avg_step_time_sec": 22.21014424406964, "gpu_allocated_gb": 32.363495424, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2555.9702155590057, "eta_sec": 1887.8622607459192, "timestamp": "2026-02-15T14:16:28.953217"}
{"step": 116, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.036328125, "mean_total_reward": 0.7963671875, "std_total_reward": 0.4003499962713442, "total_tokens": 3720, "mean_completion_length": 74.4, "median_completion_length": 72, "min_completion_length": 41, "max_completion_length": 127, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0006804766706506563, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 4.223826885223389, "avg_step_time_sec": 22.055089784079584, "gpu_allocated_gb": 32.36322304, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2560.2004764080048, "eta_sec": 1852.6275418626851, "timestamp": "2026-02-15T14:16:31.682683"}
{"step": 117, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.0980859375, "mean_total_reward": 0.79019140625, "std_total_reward": 0.40501042819313976, "total_tokens": 10044, "mean_completion_length": 200.88, "median_completion_length": 182, "min_completion_length": 61, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.0063969886365323626, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.2556333541870117, "avg_step_time_sec": 21.88586366074717, "gpu_allocated_gb": 32.362932224, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2562.461991071701, "eta_sec": 1816.526683842015, "timestamp": "2026-02-15T14:16:45.055174"}
{"step": 118, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.096513671875, "mean_total_reward": 0.6103486328125, "std_total_reward": 0.4878605904206811, "total_tokens": 9883, "mean_completion_length": 197.66, "median_completion_length": 156, "min_completion_length": 46, "max_completion_length": 663, "frac_has_think": 1.0, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.2784809312320287, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 20.713245630264282, "avg_step_time_sec": 21.875926219810875, "gpu_allocated_gb": 32.365725696, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2583.183284521103, "eta_sec": 1793.8259500244917, "timestamp": "2026-02-15T14:16:58.278745"}
{"step": 119, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.05376953125, "mean_total_reward": 0.834623046875, "std_total_reward": 0.36678038913651456, "total_tokens": 5506, "mean_completion_length": 110.12, "median_completion_length": 98, "min_completion_length": 44, "max_completion_length": 454, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.19298944660906484, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.655266761779785, "avg_step_time_sec": 21.75642487982742, "gpu_allocated_gb": 32.363750912, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2590.8447449207306, "eta_sec": 1762.270415266021, "timestamp": "2026-02-15T14:17:04.244479"}
{"step": 120, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.1351171875, "mean_total_reward": 0.56648828125, "std_total_reward": 0.506000102348468, "total_tokens": 13836, "mean_completion_length": 276.72, "median_completion_length": 95, "min_completion_length": 22, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.06748261468475245, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 5.081869840621948, "avg_step_time_sec": 21.617470254500706, "gpu_allocated_gb": 32.363410432, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2595.9324975013733, "eta_sec": 1729.3976203600564, "timestamp": "2026-02-15T14:17:20.185701"}
{"step": 121, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.109482421875, "mean_total_reward": 0.7490517578125, "std_total_reward": 0.432284152803621, "total_tokens": 11211, "mean_completion_length": 224.22, "median_completion_length": 96, "min_completion_length": 34, "max_completion_length": 969, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2762700582517344, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.224384307861328, "avg_step_time_sec": 21.622486073123522, "gpu_allocated_gb": 32.365681664, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2618.162877559662, "eta_sec": 1708.1763997767582, "timestamp": "2026-02-15T14:17:35.174456"}
{"step": 122, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.64, "mean_correctness_reward": 0.64, "mean_length_penalty": -0.0950390625, "mean_total_reward": 0.63049609375, "std_total_reward": 0.48469085752798746, "total_tokens": 9732, "mean_completion_length": 194.64, "median_completion_length": 117, "min_completion_length": 50, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.2206802323956573, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 10.607435464859009, "avg_step_time_sec": 21.53219877305578, "gpu_allocated_gb": 32.364159488, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2628.7761266231537, "eta_sec": 1679.5115042983507, "timestamp": "2026-02-15T14:17:51.834848"}
{"step": 123, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.040126953125, "mean_total_reward": 0.6959873046875, "std_total_reward": 0.4593249021125697, "total_tokens": 4109, "mean_completion_length": 82.18, "median_completion_length": 72, "min_completion_length": 25, "max_completion_length": 194, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.10035865284504406, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 20.784200191497803, "avg_step_time_sec": 21.526117483774822, "gpu_allocated_gb": 32.365642752, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2649.5666007995605, "eta_sec": 1657.5110462506614, "timestamp": "2026-02-15T14:18:01.449999"}
{"step": 124, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.039208984375, "mean_total_reward": 0.5760791015625, "std_total_reward": 0.49400484160027547, "total_tokens": 4015, "mean_completion_length": 80.3, "median_completion_length": 61, "min_completion_length": 30, "max_completion_length": 176, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.22021849831196885, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.734344005584717, "avg_step_time_sec": 21.374570923466838, "gpu_allocated_gb": 32.363029504, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2652.307273387909, "eta_sec": 1624.4673901834797, "timestamp": "2026-02-15T14:18:04.084359"}
{"step": 125, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.116943359375, "mean_total_reward": 0.4883056640625, "std_total_reward": 0.5014199909242446, "total_tokens": 11975, "mean_completion_length": 239.5, "median_completion_length": 118, "min_completion_length": 38, "max_completion_length": 2036, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.10091307042230085, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 2.568218946456909, "avg_step_time_sec": 21.224120107650755, "gpu_allocated_gb": 32.363011584, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2654.8814861774445, "eta_sec": 1591.8090080738066, "timestamp": "2026-02-15T14:18:17.782425"}
{"step": 126, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.05693359375, "mean_total_reward": 0.814306640625, "std_total_reward": 0.3834519939590125, "total_tokens": 5830, "mean_completion_length": 116.6, "median_completion_length": 95, "min_completion_length": 36, "max_completion_length": 328, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.23833510131236282, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 20.96297836303711, "avg_step_time_sec": 21.222047554122078, "gpu_allocated_gb": 32.36568064, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2675.851338148117, "eta_sec": 1570.4315190050338, "timestamp": "2026-02-15T14:18:28.641697"}
{"step": 127, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.045986328125, "mean_total_reward": 0.9354013671875, "std_total_reward": 0.24024634801091585, "total_tokens": 4709, "mean_completion_length": 94.18, "median_completion_length": 60, "min_completion_length": 36, "max_completion_length": 532, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09271646946487148, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.0397844314575195, "avg_step_time_sec": 21.086754143707395, "gpu_allocated_gb": 32.363231744, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2679.8972778320312, "eta_sec": 1539.3330524906398, "timestamp": "2026-02-15T14:18:33.659688"}
{"step": 128, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.44, "mean_correctness_reward": 0.44, "mean_length_penalty": -0.084189453125, "mean_total_reward": 0.4315810546875, "std_total_reward": 0.49922577370157617, "total_tokens": 8621, "mean_completion_length": 172.42, "median_completion_length": 105, "min_completion_length": 43, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08198600663113509, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 5.656829118728638, "avg_step_time_sec": 20.96620785444975, "gpu_allocated_gb": 32.363501568, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2685.559995174408, "eta_sec": 1509.566965520382, "timestamp": "2026-02-15T14:18:48.030468"}
{"step": 129, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.069599609375, "mean_total_reward": 0.8130400390625, "std_total_reward": 0.3911295793130536, "total_tokens": 7127, "mean_completion_length": 142.54, "median_completion_length": 76, "min_completion_length": 39, "max_completion_length": 777, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06048167037900855, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 20.706263542175293, "avg_step_time_sec": 20.964192782261573, "gpu_allocated_gb": 32.365645824, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2706.2729227542877, "eta_sec": 1488.4576875405717, "timestamp": "2026-02-15T14:19:01.799800"}
{"step": 130, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.06578125, "mean_total_reward": 0.693421875, "std_total_reward": 0.458919854802744, "total_tokens": 6736, "mean_completion_length": 134.72, "median_completion_length": 94, "min_completion_length": 20, "max_completion_length": 376, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.17240548694921348, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.294755458831787, "avg_step_time_sec": 20.866735572081346, "gpu_allocated_gb": 32.36386304, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2714.5737183094025, "eta_sec": 1460.6714900456943, "timestamp": "2026-02-15T14:19:07.747408"}
{"step": 131, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.055048828125, "mean_total_reward": 0.7344951171875, "std_total_reward": 0.43772946603893176, "total_tokens": 5637, "mean_completion_length": 112.74, "median_completion_length": 65, "min_completion_length": 31, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1398179145814682, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 4.7613725662231445, "avg_step_time_sec": 20.74379386974655, "gpu_allocated_gb": 32.3633024, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2719.3412477970123, "eta_sec": 1431.321777012512, "timestamp": "2026-02-15T14:19:21.382130"}
{"step": 132, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.048447265625, "mean_total_reward": 0.9551552734375, "std_total_reward": 0.1961245078693972, "total_tokens": 4961, "mean_completion_length": 99.22, "median_completion_length": 94, "min_completion_length": 26, "max_completion_length": 239, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12059301162355104, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 19.652993202209473, "avg_step_time_sec": 20.735530228325814, "gpu_allocated_gb": 32.365605888, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2738.9991166591644, "eta_sec": 1410.0160555261555, "timestamp": "2026-02-15T14:19:31.234638"}
{"step": 133, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.0833203125, "mean_total_reward": 0.95166796875, "std_total_reward": 0.19989598963819008, "total_tokens": 8532, "mean_completion_length": 170.64, "median_completion_length": 57, "min_completion_length": 38, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08668944031295832, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 3.2134487628936768, "avg_step_time_sec": 20.603785254901513, "gpu_allocated_gb": 32.363117568, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2742.218743085861, "eta_sec": 1380.4536120784014, "timestamp": "2026-02-15T14:19:44.904589"}
{"step": 134, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.46, "mean_correctness_reward": 0.46, "mean_length_penalty": -0.154609375, "mean_total_reward": 0.44453906249999997, "std_total_reward": 0.5053818915248948, "total_tokens": 15832, "mean_completion_length": 316.64, "median_completion_length": 131, "min_completion_length": 61, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.41091618637938293, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 20.45426869392395, "avg_step_time_sec": 20.602669459670338, "gpu_allocated_gb": 32.365641728, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2762.678447008133, "eta_sec": 1359.7761843382423, "timestamp": "2026-02-15T14:20:06.589707"}
{"step": 135, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.089931640625, "mean_total_reward": 0.5910068359375, "std_total_reward": 0.49499672471417455, "total_tokens": 9209, "mean_completion_length": 184.18, "median_completion_length": 97, "min_completion_length": 43, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.191070629700543, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 21.745208501815796, "avg_step_time_sec": 20.61113271183438, "gpu_allocated_gb": 32.365657088, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2784.4301528930664, "eta_sec": 1339.7236262692347, "timestamp": "2026-02-15T14:20:27.558039"}
{"step": 136, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.38, "mean_correctness_reward": 0.38, "mean_length_penalty": -0.052314453125, "mean_total_reward": 0.3747685546875, "std_total_reward": 0.48727054661939123, "total_tokens": 5357, "mean_completion_length": 107.14, "median_completion_length": 111, "min_completion_length": 25, "max_completion_length": 294, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.060923953367494954, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 20.805468559265137, "avg_step_time_sec": 20.612561651889017, "gpu_allocated_gb": 32.365615104, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2805.241674900055, "eta_sec": 1319.203945720897, "timestamp": "2026-02-15T14:20:37.801336"}
{"step": 137, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.071318359375, "mean_total_reward": 0.5928681640625, "std_total_reward": 0.4950731873988261, "total_tokens": 7303, "mean_completion_length": 146.06, "median_completion_length": 73, "min_completion_length": 40, "max_completion_length": 618, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0018719907361527187, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 4.282426118850708, "avg_step_time_sec": 20.493363582304795, "gpu_allocated_gb": 32.36317952, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2809.530418395996, "eta_sec": 1291.0819056852022, "timestamp": "2026-02-15T14:20:44.139739"}
{"step": 138, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.074248046875, "mean_total_reward": 0.5725751953125, "std_total_reward": 0.496723869382307, "total_tokens": 7603, "mean_completion_length": 152.06, "median_completion_length": 60, "min_completion_length": 28, "max_completion_length": 850, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.13903517467452137, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 6.850415468215942, "avg_step_time_sec": 20.39450163944908, "gpu_allocated_gb": 32.363644928, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2816.3868243694305, "eta_sec": 1264.4591016458428, "timestamp": "2026-02-15T14:20:52.390961"}
{"step": 139, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.081845703125, "mean_total_reward": 0.7118154296875, "std_total_reward": 0.456418741400944, "total_tokens": 8381, "mean_completion_length": 167.62, "median_completion_length": 71, "min_completion_length": 38, "max_completion_length": 1558, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2039682211518703, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 9.114906311035156, "avg_step_time_sec": 20.31335347161876, "gpu_allocated_gb": 32.363976192, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2825.507611513138, "eta_sec": 1239.1145617687444, "timestamp": "2026-02-15T14:21:05.559954"}
{"step": 140, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.053125, "mean_total_reward": 0.7546875, "std_total_reward": 0.4261028380311331, "total_tokens": 5440, "mean_completion_length": 108.8, "median_completion_length": 77, "min_completion_length": 21, "max_completion_length": 426, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1209632931039368, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 15.958508491516113, "avg_step_time_sec": 20.2822474360466, "gpu_allocated_gb": 32.364994048, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2841.4724049568176, "eta_sec": 1216.934846162796, "timestamp": "2026-02-15T14:21:14.819592"}
{"step": 141, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.065517578125, "mean_total_reward": 0.8334482421875, "std_total_reward": 0.3661990881533497, "total_tokens": 6709, "mean_completion_length": 134.18, "median_completion_length": 114, "min_completion_length": 22, "max_completion_length": 474, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.28045581684053694, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.947704076766968, "avg_step_time_sec": 20.17349180938504, "gpu_allocated_gb": 32.36337408, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2846.4269874095917, "eta_sec": 1190.2360167537174, "timestamp": "2026-02-15T14:21:20.117201"}
{"step": 142, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.043125, "mean_total_reward": 0.7156875, "std_total_reward": 0.4508669925419654, "total_tokens": 4416, "mean_completion_length": 88.32, "median_completion_length": 76, "min_completion_length": 16, "max_completion_length": 191, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.16052719459600787, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 5.43354868888855, "avg_step_time_sec": 20.069689393043518, "gpu_allocated_gb": 32.363436032, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2851.8663976192474, "eta_sec": 1164.041984796524, "timestamp": "2026-02-15T14:21:23.754440"}
{"step": 143, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.097841796875, "mean_total_reward": 0.5502158203124999, "std_total_reward": 0.5002293986597098, "total_tokens": 10019, "mean_completion_length": 200.38, "median_completion_length": 86, "min_completion_length": 22, "max_completion_length": 1141, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15178543065298392, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.759028196334839, "avg_step_time_sec": 19.94863581824136, "gpu_allocated_gb": 32.363024896, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2854.6314511299133, "eta_sec": 1137.0722416397575, "timestamp": "2026-02-15T14:21:32.546472"}
{"step": 144, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.035693359375, "mean_total_reward": 0.5964306640625, "std_total_reward": 0.48997915985106555, "total_tokens": 3655, "mean_completion_length": 73.1, "median_completion_length": 62, "min_completion_length": 21, "max_completion_length": 202, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.001006713882241308, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 12.08235478401184, "avg_step_time_sec": 19.894008866614765, "gpu_allocated_gb": 32.364379136, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2866.7213172912598, "eta_sec": 1114.0644965304268, "timestamp": "2026-02-15T14:21:38.617336"}
{"step": 145, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.050830078125, "mean_total_reward": 0.8349169921875, "std_total_reward": 0.3669765233487772, "total_tokens": 5205, "mean_completion_length": 104.1, "median_completion_length": 76, "min_completion_length": 24, "max_completion_length": 260, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.3028545420219507, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.771197557449341, "avg_step_time_sec": 19.775920512758454, "gpu_allocated_gb": 32.363045376, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2869.4993863105774, "eta_sec": 1087.675628201715, "timestamp": "2026-02-15T14:21:41.877014"}
{"step": 146, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.88, "mean_correctness_reward": 0.88, "mean_length_penalty": -0.05314453125, "mean_total_reward": 0.874685546875, "std_total_reward": 0.32404490352814297, "total_tokens": 5442, "mean_completion_length": 108.84, "median_completion_length": 78, "min_completion_length": 18, "max_completion_length": 291, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.23206158619779022, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 3.4472873210906982, "avg_step_time_sec": 19.664080559390865, "gpu_allocated_gb": 32.363155456, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2872.9526767730713, "eta_sec": 1061.8603502071066, "timestamp": "2026-02-15T14:21:46.019984"}
{"step": 147, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.067353515625, "mean_total_reward": 0.8332646484375, "std_total_reward": 0.36655025969847216, "total_tokens": 6897, "mean_completion_length": 137.94, "median_completion_length": 78, "min_completion_length": 19, "max_completion_length": 661, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.18029838538981813, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.179245710372925, "avg_step_time_sec": 19.558741546812513, "gpu_allocated_gb": 32.363180032, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2877.1371064186096, "eta_sec": 1036.6133019810632, "timestamp": "2026-02-15T14:21:51.959831"}
{"step": 148, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.32, "mean_correctness_reward": 0.32, "mean_length_penalty": -0.082509765625, "mean_total_reward": 0.3117490234375, "std_total_reward": 0.46879002167710315, "total_tokens": 8449, "mean_completion_length": 168.98, "median_completion_length": 169, "min_completion_length": 52, "max_completion_length": 335, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15265996853332928, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 7.194763660430908, "avg_step_time_sec": 19.47520115568831, "gpu_allocated_gb": 32.363704832, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2884.3388113975525, "eta_sec": 1012.7104600957921, "timestamp": "2026-02-15T14:21:57.391592"}
{"step": 149, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.021435546875, "mean_total_reward": 0.7178564453125, "std_total_reward": 0.4486137963642081, "total_tokens": 2195, "mean_completion_length": 43.9, "median_completion_length": 33, "min_completion_length": 19, "max_completion_length": 110, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.23150302110699497, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.448407888412476, "avg_step_time_sec": 19.37435019416297, "gpu_allocated_gb": 32.363259392, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2888.7932872772217, "eta_sec": 988.0918599023115, "timestamp": "2026-02-15T14:22:00.044015"}
{"step": 150, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.050966796875, "mean_total_reward": 0.9349033203125, "std_total_reward": 0.24033250200940853, "total_tokens": 5219, "mean_completion_length": 104.38, "median_completion_length": 60, "min_completion_length": 33, "max_completion_length": 450, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09294079929582574, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 1.9990208148956299, "avg_step_time_sec": 19.258514664967855, "gpu_allocated_gb": 32.362890752, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2890.7986233234406, "eta_sec": 962.9257332483927, "timestamp": "2026-02-15T14:22:04.877815"}
{"step": 151, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.03416015625, "mean_total_reward": 0.956583984375, "std_total_reward": 0.1969377170417483, "total_tokens": 3498, "mean_completion_length": 69.96, "median_completion_length": 43, "min_completion_length": 20, "max_completion_length": 403, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0804134063031374, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 5.110511541366577, "avg_step_time_sec": 19.164819280043343, "gpu_allocated_gb": 32.363396096, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2896.5298795700073, "eta_sec": 939.0761447221238, "timestamp": "2026-02-15T14:22:09.596298"}
{"step": 152, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.14154296875, "mean_total_reward": 0.485845703125, "std_total_reward": 0.5042779175912485, "total_tokens": 14494, "mean_completion_length": 289.88, "median_completion_length": 202, "min_completion_length": 54, "max_completion_length": 1165, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2700991340730778, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 4.763923406600952, "avg_step_time_sec": 19.070076544033853, "gpu_allocated_gb": 32.363394048, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2901.300201892853, "eta_sec": 915.3636741136249, "timestamp": "2026-02-15T14:22:19.607284"}
{"step": 153, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.05435546875, "mean_total_reward": 0.554564453125, "std_total_reward": 0.4965595820231374, "total_tokens": 5566, "mean_completion_length": 111.32, "median_completion_length": 117, "min_completion_length": 45, "max_completion_length": 201, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.3034831615180193, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 12.523290872573853, "avg_step_time_sec": 19.02728709520078, "gpu_allocated_gb": 32.364411904, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2913.829669713974, "eta_sec": 894.2824934744367, "timestamp": "2026-02-15T14:22:25.931305"}
{"step": 154, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.227705078125, "mean_total_reward": 0.7172294921875, "std_total_reward": 0.45793648165327583, "total_tokens": 23317, "mean_completion_length": 466.34, "median_completion_length": 190, "min_completion_length": 31, "max_completion_length": 2048, "frac_has_think": 0.92, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.3259426782110147, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 3.0928993225097656, "avg_step_time_sec": 18.92381704472876, "gpu_allocated_gb": 32.363104768, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2916.928559064865, "eta_sec": 870.495584057523, "timestamp": "2026-02-15T14:22:41.470297"}
{"step": 155, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.92, "mean_correctness_reward": 0.92, "mean_length_penalty": -0.13123046875, "mean_total_reward": 0.906876953125, "std_total_reward": 0.2840869674569531, "total_tokens": 13438, "mean_completion_length": 268.76, "median_completion_length": 76, "min_completion_length": 36, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.09929027451321895, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.77408719062805, "avg_step_time_sec": 18.94865749728295, "gpu_allocated_gb": 32.365646848, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2939.710252761841, "eta_sec": 852.6895873777328, "timestamp": "2026-02-15T14:23:03.064614"}
{"step": 156, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.0397265625, "mean_total_reward": 0.77602734375, "std_total_reward": 0.41336696954014773, "total_tokens": 4068, "mean_completion_length": 81.36, "median_completion_length": 63, "min_completion_length": 16, "max_completion_length": 239, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0602854316507579, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 21.156252145767212, "avg_step_time_sec": 18.962808745029644, "gpu_allocated_gb": 32.365668352, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2960.8731746673584, "eta_sec": 834.3635847813043, "timestamp": "2026-02-15T14:23:13.052034"}
{"step": 157, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.130556640625, "mean_total_reward": 0.8069443359375, "std_total_reward": 0.3942212950936925, "total_tokens": 13369, "mean_completion_length": 267.38, "median_completion_length": 59, "min_completion_length": 32, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.2755955225594936, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 3.1820671558380127, "avg_step_time_sec": 18.862294467391482, "gpu_allocated_gb": 32.363117568, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2964.060971736908, "eta_sec": 811.0786620978338, "timestamp": "2026-02-15T14:23:27.724066"}
{"step": 158, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.045478515625, "mean_total_reward": 0.8354521484375, "std_total_reward": 0.3677677160867218, "total_tokens": 4657, "mean_completion_length": 93.14, "median_completion_length": 74, "min_completion_length": 37, "max_completion_length": 311, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2182231443906613, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 21.67961049079895, "avg_step_time_sec": 18.88012558146368, "gpu_allocated_gb": 32.36568064, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2985.746315717697, "eta_sec": 792.9652744214745, "timestamp": "2026-02-15T14:23:38.399523"}
{"step": 159, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.92, "mean_correctness_reward": 0.92, "mean_length_penalty": -0.086025390625, "mean_total_reward": 0.9113974609375, "std_total_reward": 0.27204239440036154, "total_tokens": 8809, "mean_completion_length": 176.18, "median_completion_length": 188, "min_completion_length": 65, "max_completion_length": 393, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2003181161419901, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 3.956143379211426, "avg_step_time_sec": 18.786264058179075, "gpu_allocated_gb": 32.36324352, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2989.7083241939545, "eta_sec": 770.236826385342, "timestamp": "2026-02-15T14:23:43.102542"}
{"step": 160, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.189423828125, "mean_total_reward": 0.6810576171875, "std_total_reward": 0.4806528787660238, "total_tokens": 19397, "mean_completion_length": 387.94, "median_completion_length": 93, "min_completion_length": 21, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.9, "mean_reward_std_per_prompt": 0.1064141256897885, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 4.867472171783447, "avg_step_time_sec": 18.699271608889102, "gpu_allocated_gb": 32.36333056, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 2994.5804884433746, "eta_sec": 747.9708643555641, "timestamp": "2026-02-15T14:23:58.900139"}
{"step": 161, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.069970703125, "mean_total_reward": 0.7330029296875, "std_total_reward": 0.43780987585616876, "total_tokens": 7165, "mean_completion_length": 143.3, "median_completion_length": 148, "min_completion_length": 44, "max_completion_length": 389, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.18136524149119765, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.342324256896973, "avg_step_time_sec": 18.72189926508791, "gpu_allocated_gb": 32.365683712, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3016.9278059005737, "eta_sec": 730.1540713384285, "timestamp": "2026-02-15T14:24:10.127224"}
{"step": 162, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.036298828125, "mean_total_reward": 0.9563701171875, "std_total_reward": 0.19660134620525757, "total_tokens": 3717, "mean_completion_length": 74.34, "median_completion_length": 72, "min_completion_length": 35, "max_completion_length": 195, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12080978011487396, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.577805280685425, "avg_step_time_sec": 18.634590042961968, "gpu_allocated_gb": 32.363295744, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3021.510771751404, "eta_sec": 708.1144216325548, "timestamp": "2026-02-15T14:24:13.383212"}
{"step": 163, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.077744140625, "mean_total_reward": 0.5922255859375, "std_total_reward": 0.49151028144556724, "total_tokens": 7961, "mean_completion_length": 159.22, "median_completion_length": 178, "min_completion_length": 34, "max_completion_length": 260, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0010092643515935043, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 3.3329195976257324, "avg_step_time_sec": 18.540714764156224, "gpu_allocated_gb": 32.363067904, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3024.8493180274963, "eta_sec": 686.0064462737803, "timestamp": "2026-02-15T14:24:17.389747"}
{"step": 164, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.03787109375, "mean_total_reward": 0.776212890625, "std_total_reward": 0.41383371703157884, "total_tokens": 3878, "mean_completion_length": 77.56, "median_completion_length": 71, "min_completion_length": 37, "max_completion_length": 155, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06043053551078907, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 3.559296131134033, "avg_step_time_sec": 18.449364650540236, "gpu_allocated_gb": 32.363123712, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3028.4142503738403, "eta_sec": 664.1771274194485, "timestamp": "2026-02-15T14:24:20.132472"}
{"step": 165, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.86, "mean_correctness_reward": 0.86, "mean_length_penalty": -0.024072265625, "mean_total_reward": 0.8575927734375, "std_total_reward": 0.3474374738536426, "total_tokens": 2465, "mean_completion_length": 49.3, "median_completion_length": 34, "min_completion_length": 19, "max_completion_length": 127, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1896252233005497, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.569873332977295, "avg_step_time_sec": 18.35312530922167, "gpu_allocated_gb": 32.363033088, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3030.9893760681152, "eta_sec": 642.3593858227584, "timestamp": "2026-02-15T14:24:22.369645"}
{"step": 166, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.095087890625, "mean_total_reward": 0.5304912109375, "std_total_reward": 0.5032924797190916, "total_tokens": 9737, "mean_completion_length": 194.74, "median_completion_length": 140, "min_completion_length": 19, "max_completion_length": 1004, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.16046135701737524, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.1355559825897217, "avg_step_time_sec": 18.25542910845883, "gpu_allocated_gb": 32.362935296, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3033.1306829452515, "eta_sec": 620.6845896876002, "timestamp": "2026-02-15T14:24:30.212601"}
{"step": 167, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.040185546875, "mean_total_reward": 0.7559814453125, "std_total_reward": 0.427629338854355, "total_tokens": 4115, "mean_completion_length": 82.3, "median_completion_length": 54, "min_completion_length": 18, "max_completion_length": 311, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08096006537814464, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 10.822039365768433, "avg_step_time_sec": 18.210917792634337, "gpu_allocated_gb": 32.364221952, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3043.960635662079, "eta_sec": 600.9602871569331, "timestamp": "2026-02-15T14:24:36.575526"}
{"step": 168, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.38, "mean_correctness_reward": 0.38, "mean_length_penalty": -0.07162109375, "mean_total_reward": 0.372837890625, "std_total_reward": 0.48845831457928374, "total_tokens": 7334, "mean_completion_length": 146.68, "median_completion_length": 112, "min_completion_length": 23, "max_completion_length": 1250, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06358054671201796, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 3.944023370742798, "avg_step_time_sec": 18.125995802027838, "gpu_allocated_gb": 32.363246592, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3047.910614967346, "eta_sec": 580.0318656648908, "timestamp": "2026-02-15T14:24:45.941167"}
{"step": 169, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.123701171875, "mean_total_reward": 0.5676298828125, "std_total_reward": 0.5033719054203477, "total_tokens": 12667, "mean_completion_length": 253.34, "median_completion_length": 74, "min_completion_length": 45, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.2568387546099284, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 12.540778160095215, "avg_step_time_sec": 18.092947176927645, "gpu_allocated_gb": 32.36451584, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3060.458494901657, "eta_sec": 560.881362484757, "timestamp": "2026-02-15T14:25:04.080276"}
{"step": 170, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.86, "mean_correctness_reward": 0.86, "mean_length_penalty": -0.0363671875, "mean_total_reward": 0.85636328125, "std_total_reward": 0.34654324552057014, "total_tokens": 3724, "mean_completion_length": 74.48, "median_completion_length": 70, "min_completion_length": 18, "max_completion_length": 164, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.243460831745483, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 21.309808492660522, "avg_step_time_sec": 18.111869890549603, "gpu_allocated_gb": 32.365618176, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3081.7742142677307, "eta_sec": 543.3560967164881, "timestamp": "2026-02-15T14:25:13.465676"}
{"step": 171, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.03857421875, "mean_total_reward": 0.596142578125, "std_total_reward": 0.4919320044634023, "total_tokens": 3950, "mean_completion_length": 79, "median_completion_length": 53, "min_completion_length": 20, "max_completion_length": 264, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.000589706440265408, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 2.4594829082489014, "avg_step_time_sec": 18.020335463752524, "gpu_allocated_gb": 32.362997248, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3084.24054980278, "eta_sec": 522.5897284488232, "timestamp": "2026-02-15T14:25:16.616669"}
{"step": 172, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.058623046875, "mean_total_reward": 0.6541376953125, "std_total_reward": 0.47556709496590294, "total_tokens": 6003, "mean_completion_length": 120.06, "median_completion_length": 97, "min_completion_length": 19, "max_completion_length": 728, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2133622725826351, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 3.6800951957702637, "avg_step_time_sec": 17.936961973822395, "gpu_allocated_gb": 32.363204608, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3087.9261190891266, "eta_sec": 502.2349352670271, "timestamp": "2026-02-15T14:25:22.956887"}
{"step": 173, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.0275, "mean_total_reward": 0.77725, "std_total_reward": 0.413816321648744, "total_tokens": 2816, "mean_completion_length": 56.32, "median_completion_length": 64, "min_completion_length": 12, "max_completion_length": 124, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.060448528664057845, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.147762775421143, "avg_step_time_sec": 17.880377007357648, "gpu_allocated_gb": 32.363792896, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3096.0801463127136, "eta_sec": 482.7701791986565, "timestamp": "2026-02-15T14:25:27.469754"}
{"step": 174, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.04837890625, "mean_total_reward": 0.775162109375, "std_total_reward": 0.41363394633094613, "total_tokens": 4954, "mean_completion_length": 99.08, "median_completion_length": 73, "min_completion_length": 26, "max_completion_length": 270, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.31809537879767025, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.143486738204956, "avg_step_time_sec": 17.78993510925907, "gpu_allocated_gb": 32.362927616, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3098.229461669922, "eta_sec": 462.5383128407358, "timestamp": "2026-02-15T14:25:30.638904"}
{"step": 175, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.0466796875, "mean_total_reward": 0.71533203125, "std_total_reward": 0.4482150384530796, "total_tokens": 4780, "mean_completion_length": 95.6, "median_completion_length": 42, "min_completion_length": 21, "max_completion_length": 334, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.20019791286758074, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 3.4469797611236572, "avg_step_time_sec": 17.70797536441258, "gpu_allocated_gb": 32.363147264, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3101.6817820072174, "eta_sec": 442.6993841103145, "timestamp": "2026-02-15T14:25:34.466153"}
{"step": 176, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.02876953125, "mean_total_reward": 0.7171230468750001, "std_total_reward": 0.4492831972749562, "total_tokens": 2946, "mean_completion_length": 58.92, "median_completion_length": 63, "min_completion_length": 17, "max_completion_length": 117, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.31164960959040533, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.056007385253906, "avg_step_time_sec": 17.630407364530996, "gpu_allocated_gb": 32.363239936, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3105.7442059516907, "eta_sec": 423.1297767487439, "timestamp": "2026-02-15T14:25:37.085747"}
{"step": 177, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.019033203125, "mean_total_reward": 0.7380966796875, "std_total_reward": 0.43899047980815054, "total_tokens": 1949, "mean_completion_length": 38.98, "median_completion_length": 25, "min_completion_length": 13, "max_completion_length": 104, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09192178617723784, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.0992722511291504, "avg_step_time_sec": 17.54266083846658, "gpu_allocated_gb": 32.362912768, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3107.848968744278, "eta_sec": 403.48119928473136, "timestamp": "2026-02-15T14:25:39.081085"}
{"step": 178, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.68, "mean_correctness_reward": 0.68, "mean_length_penalty": -0.029873046875, "mean_total_reward": 0.6770126953125, "std_total_reward": 0.4655885160646006, "total_tokens": 3059, "mean_completion_length": 61.18, "median_completion_length": 65, "min_completion_length": 19, "max_completion_length": 159, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.21790177251613893, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 1.9720134735107422, "avg_step_time_sec": 17.455185291472446, "gpu_allocated_gb": 32.362922496, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3109.8274533748627, "eta_sec": 384.0140764123938, "timestamp": "2026-02-15T14:25:41.470335"}
{"step": 179, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.52, "mean_correctness_reward": 0.52, "mean_length_penalty": -0.148154296875, "mean_total_reward": 0.5051845703125, "std_total_reward": 0.5127541831611726, "total_tokens": 15171, "mean_completion_length": 303.42, "median_completion_length": 98, "min_completion_length": 33, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.10372623244695044, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 2.4071593284606934, "avg_step_time_sec": 17.371118107321543, "gpu_allocated_gb": 32.363003392, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3112.242645263672, "eta_sec": 364.7934802537524, "timestamp": "2026-02-15T14:25:56.193907"}
{"step": 180, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.094150390625, "mean_total_reward": 0.5305849609375, "std_total_reward": 0.505430557896712, "total_tokens": 9641, "mean_completion_length": 192.82, "median_completion_length": 48, "min_completion_length": 20, "max_completion_length": 1500, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.18368317255483035, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.08407688140869, "avg_step_time_sec": 17.397301211622025, "gpu_allocated_gb": 32.365698048, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3134.3329417705536, "eta_sec": 347.9460242324405, "timestamp": "2026-02-15T14:26:14.132067"}
{"step": 181, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.0358203125, "mean_total_reward": 0.49641796875, "std_total_reward": 0.4996860444258244, "total_tokens": 3668, "mean_completion_length": 73.36, "median_completion_length": 60, "min_completion_length": 31, "max_completion_length": 160, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.22029246587602017, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 15.368210792541504, "avg_step_time_sec": 17.386090767317715, "gpu_allocated_gb": 32.36488704, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3149.707567214966, "eta_sec": 330.3357245790366, "timestamp": "2026-02-15T14:26:21.368192"}
{"step": 182, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.02841796875, "mean_total_reward": 0.937158203125, "std_total_reward": 0.23791647203463082, "total_tokens": 2910, "mean_completion_length": 58.2, "median_completion_length": 52, "min_completion_length": 21, "max_completion_length": 153, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.14016522054520822, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.4553351402282715, "avg_step_time_sec": 17.304053648487553, "gpu_allocated_gb": 32.362988032, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3152.1687903404236, "eta_sec": 311.47296567277596, "timestamp": "2026-02-15T14:26:24.268544"}
{"step": 183, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.045693359375, "mean_total_reward": 0.5954306640625, "std_total_reward": 0.49137514127769566, "total_tokens": 4679, "mean_completion_length": 93.58, "median_completion_length": 41, "min_completion_length": 16, "max_completion_length": 433, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.23152607020334023, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.8524563312530518, "avg_step_time_sec": 17.225083171344195, "gpu_allocated_gb": 32.36296704, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3155.0279247760773, "eta_sec": 292.8264139128513, "timestamp": "2026-02-15T14:26:28.368156"}
{"step": 184, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.34, "mean_correctness_reward": 0.34, "mean_length_penalty": -0.0840625, "mean_total_reward": 0.33159375, "std_total_reward": 0.47819881103444395, "total_tokens": 8608, "mean_completion_length": 172.16, "median_completion_length": 54, "min_completion_length": 23, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.24318759745102714, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 4.915210723876953, "avg_step_time_sec": 17.15818169065144, "gpu_allocated_gb": 32.36337664, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3159.9490463733673, "eta_sec": 274.530907050423, "timestamp": "2026-02-15T14:26:42.765214"}
{"step": 185, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.34, "mean_correctness_reward": 0.34, "mean_length_penalty": -0.050263671875, "mean_total_reward": 0.3349736328125, "std_total_reward": 0.4741402994398331, "total_tokens": 5147, "mean_completion_length": 102.94, "median_completion_length": 58, "min_completion_length": 28, "max_completion_length": 721, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09335354408220518, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 20.661183834075928, "avg_step_time_sec": 17.17711683737265, "gpu_allocated_gb": 32.365664256, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3180.615620613098, "eta_sec": 257.6567525605898, "timestamp": "2026-02-15T14:26:55.579914"}
{"step": 186, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.04271484375, "mean_total_reward": 0.815728515625, "std_total_reward": 0.3842246290429622, "total_tokens": 4374, "mean_completion_length": 87.48, "median_completion_length": 83, "min_completion_length": 40, "max_completion_length": 142, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.23810753599198595, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.844021797180176, "avg_step_time_sec": 17.126938907049034, "gpu_allocated_gb": 32.363838464, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3188.4661712646484, "eta_sec": 239.77714469868647, "timestamp": "2026-02-15T14:27:00.099136"}
{"step": 187, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.02892578125, "mean_total_reward": 0.937107421875, "std_total_reward": 0.2375282011776187, "total_tokens": 2962, "mean_completion_length": 59.24, "median_completion_length": 51, "min_completion_length": 35, "max_completion_length": 156, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.14021296470295114, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.4065558910369873, "avg_step_time_sec": 17.04822028129496, "gpu_allocated_gb": 32.362985472, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3190.878487586975, "eta_sec": 221.62686365683447, "timestamp": "2026-02-15T14:27:02.558746"}
{"step": 188, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.078642578125, "mean_total_reward": 0.6921357421875, "std_total_reward": 0.462991939318998, "total_tokens": 8053, "mean_completion_length": 161.06, "median_completion_length": 80, "min_completion_length": 40, "max_completion_length": 1086, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2388614542384386, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.58063006401062, "avg_step_time_sec": 16.971265014181746, "gpu_allocated_gb": 32.363057664, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3193.464821100235, "eta_sec": 203.65518017018096, "timestamp": "2026-02-15T14:27:10.716961"}
{"step": 189, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.034287109375, "mean_total_reward": 0.9365712890625, "std_total_reward": 0.2383028025381099, "total_tokens": 3511, "mean_completion_length": 70.22, "median_completion_length": 51, "min_completion_length": 24, "max_completion_length": 229, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09171012140021935, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 11.250971555709839, "avg_step_time_sec": 16.940998911226867, "gpu_allocated_gb": 32.364312064, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3204.7214810848236, "eta_sec": 186.35098802349555, "timestamp": "2026-02-15T14:27:16.770445"}
{"step": 190, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.031083984375, "mean_total_reward": 0.6968916015625, "std_total_reward": 0.45906765010949346, "total_tokens": 3183, "mean_completion_length": 63.66, "median_completion_length": 54, "min_completion_length": 24, "max_completion_length": 276, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.28774304479464485, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.9472763538360596, "avg_step_time_sec": 16.86734773987218, "gpu_allocated_gb": 32.363072512, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3207.6744565963745, "eta_sec": 168.6734773987218, "timestamp": "2026-02-15T14:27:19.951613"}
{"step": 191, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.114638671875, "mean_total_reward": 0.3885361328125, "std_total_reward": 0.4968584711255009, "total_tokens": 11739, "mean_completion_length": 234.78, "median_completion_length": 143, "min_completion_length": 33, "max_completion_length": 908, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.005935542762270167, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 3.3567256927490234, "avg_step_time_sec": 16.796611498787765, "gpu_allocated_gb": 32.363158528, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3211.036760568619, "eta_sec": 151.1695034890899, "timestamp": "2026-02-15T14:27:27.844653"}
{"step": 192, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.04005859375, "mean_total_reward": 0.555994140625, "std_total_reward": 0.49503527742544867, "total_tokens": 4102, "mean_completion_length": 82.04, "median_completion_length": 49, "min_completion_length": 15, "max_completion_length": 257, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08034131397696245, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 11.011376857757568, "avg_step_time_sec": 16.766480068365734, "gpu_allocated_gb": 32.364156416, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3222.0555880069733, "eta_sec": 134.13184054692587, "timestamp": "2026-02-15T14:27:34.311228"}
{"step": 193, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.07744140625, "mean_total_reward": 0.732255859375, "std_total_reward": 0.44760484277101403, "total_tokens": 7930, "mean_completion_length": 158.6, "median_completion_length": 56, "min_completion_length": 18, "max_completion_length": 2022, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.14675799092105746, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 3.2532448768615723, "avg_step_time_sec": 16.69646330571545, "gpu_allocated_gb": 32.363117056, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3225.3139474391937, "eta_sec": 116.87524314000815, "timestamp": "2026-02-15T14:27:48.291521"}
{"step": 194, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.46, "mean_correctness_reward": 0.46, "mean_length_penalty": -0.02193359375, "mean_total_reward": 0.457806640625, "std_total_reward": 0.4982539653170485, "total_tokens": 2246, "mean_completion_length": 44.92, "median_completion_length": 39, "min_completion_length": 20, "max_completion_length": 180, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15835680189715076, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 21.002349376678467, "avg_step_time_sec": 16.718658594741036, "gpu_allocated_gb": 32.365676544, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3246.3212065696716, "eta_sec": 100.31195156844622, "timestamp": "2026-02-15T14:27:57.955270"}
{"step": 195, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.043994140625, "mean_total_reward": 0.5756005859375, "std_total_reward": 0.4958539434712187, "total_tokens": 4505, "mean_completion_length": 90.1, "median_completion_length": 58, "min_completion_length": 38, "max_completion_length": 607, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06217648739176533, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 2.4398226737976074, "avg_step_time_sec": 16.645433795146452, "gpu_allocated_gb": 32.36303104, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3248.766069173813, "eta_sec": 83.22716897573227, "timestamp": "2026-02-15T14:28:02.925963"}
{"step": 196, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.096728515625, "mean_total_reward": 0.7903271484375, "std_total_reward": 0.4070369453544248, "total_tokens": 9905, "mean_completion_length": 198.1, "median_completion_length": 98, "min_completion_length": 41, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.18521651307501613, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.299717664718628, "avg_step_time_sec": 16.59264952917488, "gpu_allocated_gb": 32.36361216, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3255.070477247238, "eta_sec": 66.37059811669953, "timestamp": "2026-02-15T14:28:17.860955"}
{"step": 197, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.02783203125, "mean_total_reward": 0.777216796875, "std_total_reward": 0.4152336904640333, "total_tokens": 2850, "mean_completion_length": 57, "median_completion_length": 54, "min_completion_length": 23, "max_completion_length": 194, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0606301872108954, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 20.560688257217407, "avg_step_time_sec": 16.612791857743627, "gpu_allocated_gb": 32.365645824, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3275.6360371112823, "eta_sec": 49.83837557323088, "timestamp": "2026-02-15T14:28:27.624385"}
{"step": 198, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.022216796875, "mean_total_reward": 0.9777783203125, "std_total_reward": 0.14005578811339875, "total_tokens": 2275, "mean_completion_length": 45.5, "median_completion_length": 40, "min_completion_length": 19, "max_completion_length": 94, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06011410851564947, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 3.4282875061035156, "avg_step_time_sec": 16.546203451927262, "gpu_allocated_gb": 32.363230208, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3279.069963693619, "eta_sec": 33.092406903854524, "timestamp": "2026-02-15T14:28:30.249872"}
{"step": 199, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.34, "mean_correctness_reward": 0.34, "mean_length_penalty": -0.1632421875, "mean_total_reward": 0.32367578125, "std_total_reward": 0.4838822793257818, "total_tokens": 16716, "mean_completion_length": 334.32, "median_completion_length": 87, "min_completion_length": 26, "max_completion_length": 1824, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09887768990995782, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 1.9095375537872314, "avg_step_time_sec": 16.472652367011985, "gpu_allocated_gb": 32.362895872, "gpu_reserved_gb": 79.958114304, "gpu_peak_gb": 74.53288704, "elapsed_sec": 3280.9843134880066, "eta_sec": 16.472652367011985, "timestamp": "2026-02-15T14:28:43.456086"}