brikdavies's picture
Add gdpo_lambda0.03_from_lambda0.1_ckpt100/logs
993a168 verified
Raw
History Blame Contribute Delete
163 kB
{"step": 0, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.92, "mean_correctness_reward": 0.92, "mean_length_penalty": -0.1266015625, "mean_total_reward": 0.916201953125, "std_total_reward": 0.27169470565672404, "total_tokens": 12964, "mean_completion_length": 259.28, "median_completion_length": 253, "min_completion_length": 107, "max_completion_length": 567, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15207345222886257, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "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": "2026-02-15T16:56:37.787406"}
{"step": 1, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.201171875, "mean_total_reward": 0.59396484375, "std_total_reward": 0.4921357779859603, "total_tokens": 20600, "mean_completion_length": 412, "median_completion_length": 328, "min_completion_length": 89, "max_completion_length": 1107, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0011135300444176277, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 7.153948068618774, "avg_step_time_sec": 7.153948068618774, "gpu_allocated_gb": 32.437665792, "gpu_reserved_gb": 46.403682304, "gpu_peak_gb": 44.284534272, "elapsed_sec": 7.159990310668945, "eta_sec": 1423.635665655136, "timestamp": "2026-02-15T16:56:48.801955"}
{"step": 2, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.149697265625, "mean_total_reward": 0.57550908203125, "std_total_reward": 0.49279035677397115, "total_tokens": 15329, "mean_completion_length": 306.58, "median_completion_length": 269, "min_completion_length": 188, "max_completion_length": 672, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06053143667026187, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 12.886689901351929, "avg_step_time_sec": 10.020318984985352, "gpu_allocated_gb": 32.438431232, "gpu_reserved_gb": 59.049508864, "gpu_peak_gb": 55.272797184, "elapsed_sec": 20.054543018341064, "eta_sec": 1984.0231590270996, "timestamp": "2026-02-15T16:56:58.563624"}
{"step": 3, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.21736328125, "mean_total_reward": 0.7934791015625, "std_total_reward": 0.4028908374989119, "total_tokens": 22258, "mean_completion_length": 445.16, "median_completion_length": 346, "min_completion_length": 205, "max_completion_length": 1016, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0008162403688615537, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.203749895095825, "avg_step_time_sec": 9.414795955022177, "gpu_allocated_gb": 32.437841408, "gpu_reserved_gb": 59.049508864, "gpu_peak_gb": 55.272797184, "elapsed_sec": 28.265708446502686, "eta_sec": 1854.7148031393688, "timestamp": "2026-02-15T16:57:09.459544"}
{"step": 4, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.14203125, "mean_total_reward": 0.9757390625, "std_total_reward": 0.14005474559611433, "total_tokens": 14544, "mean_completion_length": 290.88, "median_completion_length": 271, "min_completion_length": 127, "max_completion_length": 601, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06066465900155145, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 12.29810357093811, "avg_step_time_sec": 10.13562285900116, "gpu_allocated_gb": 32.438318592, "gpu_reserved_gb": 59.049508864, "gpu_peak_gb": 55.272797184, "elapsed_sec": 40.57225704193115, "eta_sec": 1986.5820803642273, "timestamp": "2026-02-15T16:57:18.936893"}
{"step": 5, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.237080078125, "mean_total_reward": 0.95288759765625, "std_total_reward": 0.1963542965050252, "total_tokens": 24277, "mean_completion_length": 485.54, "median_completion_length": 384, "min_completion_length": 220, "max_completion_length": 1064, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12010586683346015, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.800609111785889, "avg_step_time_sec": 9.668620109558105, "gpu_allocated_gb": 32.437737984, "gpu_reserved_gb": 59.049508864, "gpu_peak_gb": 55.272797184, "elapsed_sec": 48.38147187232971, "eta_sec": 1885.3809213638306, "timestamp": "2026-02-15T16:57:29.625817"}
{"step": 6, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.335234375, "mean_total_reward": 0.68994296875, "std_total_reward": 0.4647044932502199, "total_tokens": 34328, "mean_completion_length": 686.56, "median_completion_length": 437, "min_completion_length": 212, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.15908072102568305, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 12.390165328979492, "avg_step_time_sec": 10.12221097946167, "gpu_allocated_gb": 32.438387712, "gpu_reserved_gb": 59.049508864, "gpu_peak_gb": 55.272797184, "elapsed_sec": 60.78068733215332, "eta_sec": 1963.708930015564, "timestamp": "2026-02-15T16:57:48.850833"}
{"step": 7, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.286083984375, "mean_total_reward": 0.47141748046875, "std_total_reward": 0.5015392327211434, "total_tokens": 29295, "mean_completion_length": 585.9, "median_completion_length": 442, "min_completion_length": 210, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.10019706703761795, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 23.194334030151367, "avg_step_time_sec": 11.989657129560198, "gpu_allocated_gb": 32.439779328, "gpu_reserved_gb": 80.964747264, "gpu_peak_gb": 74.210702848, "elapsed_sec": 83.98360514640808, "eta_sec": 2314.0038260051183, "timestamp": "2026-02-15T16:58:11.641504"}
{"step": 8, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.164130859375, "mean_total_reward": 0.97507607421875, "std_total_reward": 0.13986271383784155, "total_tokens": 16807, "mean_completion_length": 336.14, "median_completion_length": 295, "min_completion_length": 216, "max_completion_length": 587, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06046339350810525, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.044317722320557, "avg_step_time_sec": 13.371489703655243, "gpu_allocated_gb": 32.439820288, "gpu_reserved_gb": 81.027661824, "gpu_peak_gb": 74.305724416, "elapsed_sec": 107.03702163696289, "eta_sec": 2567.3260231018066, "timestamp": "2026-02-15T16:58:24.897267"}
{"step": 9, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.30041015625, "mean_total_reward": 0.5909876953125, "std_total_reward": 0.4920449081245818, "total_tokens": 30762, "mean_completion_length": 615.24, "median_completion_length": 379, "min_completion_length": 118, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.20080022927162258, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 7.63866114616394, "avg_step_time_sec": 12.734508752822876, "gpu_allocated_gb": 32.437700608, "gpu_reserved_gb": 81.027661824, "gpu_peak_gb": 74.305724416, "elapsed_sec": 114.6842577457428, "eta_sec": 2432.2911717891693, "timestamp": "2026-02-15T16:58:42.478047"}
{"step": 10, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.15091796875, "mean_total_reward": 0.5554724609375, "std_total_reward": 0.49756957658308354, "total_tokens": 15454, "mean_completion_length": 309.08, "median_completion_length": 257, "min_completion_length": 112, "max_completion_length": 712, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08035589090991352, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 23.130414247512817, "avg_step_time_sec": 13.77409930229187, "gpu_allocated_gb": 32.439822336, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 137.82282638549805, "eta_sec": 2617.0788674354553, "timestamp": "2026-02-15T16:58:56.357880"}
{"step": 11, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.264111328125, "mean_total_reward": 0.79207666015625, "std_total_reward": 0.4017402410755112, "total_tokens": 27045, "mean_completion_length": 540.9, "median_completion_length": 389, "min_completion_length": 236, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.2710092282290444, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 8.394290208816528, "avg_step_time_sec": 13.285025748339566, "gpu_allocated_gb": 32.437897728, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 146.22502994537354, "eta_sec": 2510.8698664361777, "timestamp": "2026-02-15T16:59:13.380584"}
{"step": 12, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.372021484375, "mean_total_reward": 0.58883935546875, "std_total_reward": 0.4996997959184354, "total_tokens": 38095, "mean_completion_length": 761.9, "median_completion_length": 280, "min_completion_length": 102, "max_completion_length": 2048, "frac_has_think": 0.84, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.00142066001929734, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 22.181885480880737, "avg_step_time_sec": 14.02643072605133, "gpu_allocated_gb": 32.439763968, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 168.4149363040924, "eta_sec": 2636.96897649765, "timestamp": "2026-02-15T16:59:36.954229"}
{"step": 13, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.1680078125, "mean_total_reward": 0.974959765625, "std_total_reward": 0.1398433649411489, "total_tokens": 17204, "mean_completion_length": 344.08, "median_completion_length": 377, "min_completion_length": 161, "max_completion_length": 497, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06044840579546501, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 24.137807369232178, "avg_step_time_sec": 14.804228929372934, "gpu_allocated_gb": 32.439810048, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 192.5603015422821, "eta_sec": 2768.3908097927388, "timestamp": "2026-02-15T16:59:49.908825"}
{"step": 14, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.156083984375, "mean_total_reward": 0.95531748046875, "std_total_reward": 0.19563155748995095, "total_tokens": 15983, "mean_completion_length": 319.66, "median_completion_length": 337, "min_completion_length": 124, "max_completion_length": 706, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08070516159639765, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.349256277084351, "avg_step_time_sec": 14.20030231135232, "gpu_allocated_gb": 32.437569024, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 198.918066740036, "eta_sec": 2641.2562299115316, "timestamp": "2026-02-15T16:59:57.331017"}
{"step": 15, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.86, "mean_correctness_reward": 0.86, "mean_length_penalty": -0.313720703125, "mean_total_reward": 0.85058837890625, "std_total_reward": 0.35262714261773537, "total_tokens": 32125, "mean_completion_length": 642.5, "median_completion_length": 341, "min_completion_length": 178, "max_completion_length": 2048, "frac_has_think": 0.9, "frac_has_boxed": 0.9, "mean_reward_std_per_prompt": 0.22067013527075266, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 8.167296171188354, "avg_step_time_sec": 13.798101902008057, "gpu_allocated_gb": 32.437865472, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 207.0930483341217, "eta_sec": 2552.6488518714905, "timestamp": "2026-02-15T17:00:15.105294"}
{"step": 16, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.36564453125, "mean_total_reward": 0.6490306640625, "std_total_reward": 0.47762924771514353, "total_tokens": 37442, "mean_completion_length": 748.84, "median_completion_length": 634, "min_completion_length": 153, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.2885888865125658, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.247416257858276, "avg_step_time_sec": 14.388684049248695, "gpu_allocated_gb": 32.439801856, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 230.34850645065308, "eta_sec": 2647.51786506176, "timestamp": "2026-02-15T17:00:38.717620"}
{"step": 17, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.153486328125, "mean_total_reward": 0.97539541015625, "std_total_reward": 0.13983570160616832, "total_tokens": 15717, "mean_completion_length": 314.34, "median_completion_length": 311, "min_completion_length": 160, "max_completion_length": 677, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.060490975956436197, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.67157006263733, "avg_step_time_sec": 14.93473616768332, "gpu_allocated_gb": 32.439811072, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 254.028981924057, "eta_sec": 2733.0567186860476, "timestamp": "2026-02-15T17:00:52.307790"}
{"step": 18, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.209267578125, "mean_total_reward": 0.99372197265625, "std_total_reward": 0.0033450964681972957, "total_tokens": 21429, "mean_completion_length": 428.58, "median_completion_length": 417, "min_completion_length": 94, "max_completion_length": 993, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0012266083399764224, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 8.394184112548828, "avg_step_time_sec": 14.571372164620293, "gpu_allocated_gb": 32.437872128, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 262.4322111606598, "eta_sec": 2651.9897339608933, "timestamp": "2026-02-15T17:01:03.005590"}
{"step": 19, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.281572265625, "mean_total_reward": 0.83155283203125, "std_total_reward": 0.3728063484208648, "total_tokens": 28833, "mean_completion_length": 576.66, "median_completion_length": 338, "min_completion_length": 119, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15290525271952085, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 11.430001974105835, "avg_step_time_sec": 14.406036891435322, "gpu_allocated_gb": 32.43829248, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 273.8709092140198, "eta_sec": 2607.492677349793, "timestamp": "2026-02-15T17:01:21.485515"}
{"step": 20, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.2740234375, "mean_total_reward": 0.391779296875, "std_total_reward": 0.4933181108469552, "total_tokens": 28060, "mean_completion_length": 561.2, "median_completion_length": 265, "min_completion_length": 189, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.00188036860634854, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 22.655629873275757, "avg_step_time_sec": 14.818516540527344, "gpu_allocated_gb": 32.439776256, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 296.5343906879425, "eta_sec": 2667.332977294922, "timestamp": "2026-02-15T17:01:44.175804"}
{"step": 21, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.1971484375, "mean_total_reward": 0.814085546875, "std_total_reward": 0.3838178346229823, "total_tokens": 20188, "mean_completion_length": 403.76, "median_completion_length": 251, "min_completion_length": 85, "max_completion_length": 1624, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06109369042841521, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.82847762107849, "avg_step_time_sec": 15.199943258648826, "gpu_allocated_gb": 32.439785472, "gpu_reserved_gb": 81.174462464, "gpu_peak_gb": 74.30938112, "elapsed_sec": 319.37221670150757, "eta_sec": 2720.7898432981397, "timestamp": "2026-02-15T17:02:03.668290"}
{"step": 22, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.3824609375, "mean_total_reward": 0.588526171875, "std_total_reward": 0.4921881146443335, "total_tokens": 39164, "mean_completion_length": 783.28, "median_completion_length": 601, "min_completion_length": 209, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.12046113148486783, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 17.708571434020996, "avg_step_time_sec": 15.313971812074834, "gpu_allocated_gb": 32.439235072, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 337.08973693847656, "eta_sec": 2725.8869825493207, "timestamp": "2026-02-15T17:02:26.104911"}
{"step": 23, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.135625, "mean_total_reward": 0.73593125, "std_total_reward": 0.43880271052138003, "total_tokens": 13888, "mean_completion_length": 277.76, "median_completion_length": 227, "min_completion_length": 127, "max_completion_length": 902, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2749478213114635, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 24.092867374420166, "avg_step_time_sec": 15.695662923481153, "gpu_allocated_gb": 32.439754752, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 361.1919205188751, "eta_sec": 2778.1323374561644, "timestamp": "2026-02-15T17:02:40.553960"}
{"step": 24, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.10703125, "mean_total_reward": 0.9367890625, "std_total_reward": 0.23783272335368125, "total_tokens": 10960, "mean_completion_length": 219.2, "median_completion_length": 257, "min_completion_length": 68, "max_completion_length": 335, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09202623957648244, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 9.830453395843506, "avg_step_time_sec": 15.451279193162918, "gpu_allocated_gb": 32.438143488, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 371.03154730796814, "eta_sec": 2719.4251379966736, "timestamp": "2026-02-15T17:02:47.103922"}
{"step": 25, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.20458984375, "mean_total_reward": 0.7338623046875, "std_total_reward": 0.4401980782434405, "total_tokens": 20950, "mean_completion_length": 419, "median_completion_length": 427, "min_completion_length": 133, "max_completion_length": 750, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09253560899917064, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 4.645532846450806, "avg_step_time_sec": 15.019049339294433, "gpu_allocated_gb": 32.437355008, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 375.6868517398834, "eta_sec": 2628.333634376526, "timestamp": "2026-02-15T17:02:54.902353"}
{"step": 26, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.22435546875, "mean_total_reward": 0.9532693359375, "std_total_reward": 0.197127184948662, "total_tokens": 22974, "mean_completion_length": 459.48, "median_completion_length": 445, "min_completion_length": 67, "max_completion_length": 1053, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12062107150047247, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 9.195671558380127, "avg_step_time_sec": 14.79507327079773, "gpu_allocated_gb": 32.43794688, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 384.8902475833893, "eta_sec": 2574.342749118805, "timestamp": "2026-02-15T17:03:06.071226"}
{"step": 27, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.173642578125, "mean_total_reward": 0.97479072265625, "std_total_reward": 0.14014366405715625, "total_tokens": 17781, "mean_completion_length": 355.62, "median_completion_length": 288, "min_completion_length": 162, "max_completion_length": 802, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06008040537113566, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 12.167577743530273, "avg_step_time_sec": 14.697758621639675, "gpu_allocated_gb": 32.438355968, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 397.06526923179626, "eta_sec": 2542.712241543664, "timestamp": "2026-02-15T17:03:16.651083"}
{"step": 28, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.301435546875, "mean_total_reward": 0.81095693359375, "std_total_reward": 0.39400956008114213, "total_tokens": 30867, "mean_completion_length": 617.34, "median_completion_length": 335, "min_completion_length": 140, "max_completion_length": 2048, "frac_has_think": 0.82, "frac_has_boxed": 0.82, "mean_reward_std_per_prompt": 0.061564232751962863, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 9.61754584312439, "avg_step_time_sec": 14.516322450978416, "gpu_allocated_gb": 32.437993472, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 406.691721200943, "eta_sec": 2496.8074615682876, "timestamp": "2026-02-15T17:03:34.642725"}
{"step": 29, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.174638671875, "mean_total_reward": 0.81476083984375, "std_total_reward": 0.38554147563580543, "total_tokens": 17883, "mean_completion_length": 357.66, "median_completion_length": 325, "min_completion_length": 205, "max_completion_length": 670, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0605360047954809, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.986273765563965, "avg_step_time_sec": 14.808389737688262, "gpu_allocated_gb": 32.439769088, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 429.6858742237091, "eta_sec": 2532.234645144693, "timestamp": "2026-02-15T17:03:48.106522"}
{"step": 30, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.190634765625, "mean_total_reward": 0.93428095703125, "std_total_reward": 0.23698132399280117, "total_tokens": 19521, "mean_completion_length": 390.42, "median_completion_length": 306, "min_completion_length": 232, "max_completion_length": 901, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09208514904534075, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 8.026243209838867, "avg_step_time_sec": 14.582318186759949, "gpu_allocated_gb": 32.43782144, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 437.72162318229675, "eta_sec": 2478.9940917491913, "timestamp": "2026-02-15T17:03:57.624891"}
{"step": 31, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.17021484375, "mean_total_reward": 0.7948935546875, "std_total_reward": 0.40061043026001536, "total_tokens": 17430, "mean_completion_length": 348.6, "median_completion_length": 372, "min_completion_length": 184, "max_completion_length": 491, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.000612143650003724, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 10.381577968597412, "avg_step_time_sec": 14.446810437786963, "gpu_allocated_gb": 32.438158336, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 448.11136507987976, "eta_sec": 2441.5109639859966, "timestamp": "2026-02-15T17:04:05.437246"}
{"step": 32, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.295244140625, "mean_total_reward": 0.79114267578125, "std_total_reward": 0.4045943387841021, "total_tokens": 30233, "mean_completion_length": 604.66, "median_completion_length": 566, "min_completion_length": 153, "max_completion_length": 1702, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0016867936833343034, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 6.357618570327759, "avg_step_time_sec": 14.194023191928864, "gpu_allocated_gb": 32.437561344, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 454.4782304763794, "eta_sec": 2384.595896244049, "timestamp": "2026-02-15T17:04:20.620920"}
{"step": 33, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.13779296875, "mean_total_reward": 0.9558662109375, "std_total_reward": 0.19684602529298126, "total_tokens": 14110, "mean_completion_length": 282.2, "median_completion_length": 184, "min_completion_length": 75, "max_completion_length": 743, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08041215583468385, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 20.274139881134033, "avg_step_time_sec": 14.378269152207809, "gpu_allocated_gb": 32.439388672, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 474.76032638549805, "eta_sec": 2401.170948418704, "timestamp": "2026-02-15T17:04:33.432657"}
{"step": 34, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.242158203125, "mean_total_reward": 0.79273525390625, "std_total_reward": 0.40266180149488273, "total_tokens": 24797, "mean_completion_length": 495.94, "median_completion_length": 441, "min_completion_length": 237, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.0020267304960698723, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.748255729675293, "avg_step_time_sec": 14.212680522133322, "gpu_allocated_gb": 32.437985792, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 483.5174651145935, "eta_sec": 2359.3049666741313, "timestamp": "2026-02-15T17:04:50.284390"}
{"step": 35, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.133095703125, "mean_total_reward": 0.99600712890625, "std_total_reward": 0.0013724392749297702, "total_tokens": 13629, "mean_completion_length": 272.58, "median_completion_length": 246, "min_completion_length": 109, "max_completion_length": 509, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0005447615296733984, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 21.75197958946228, "avg_step_time_sec": 14.428089066914149, "gpu_allocated_gb": 32.439753728, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 505.27820563316345, "eta_sec": 2380.6346960408346, "timestamp": "2026-02-15T17:05:02.533237"}
{"step": 36, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.279375, "mean_total_reward": 0.77161875, "std_total_reward": 0.4149903448202399, "total_tokens": 28608, "mean_completion_length": 572.16, "median_completion_length": 457, "min_completion_length": 213, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.06146134703449753, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 6.591853141784668, "avg_step_time_sec": 14.210415846771664, "gpu_allocated_gb": 32.437673472, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 511.8780674934387, "eta_sec": 2330.508198870553, "timestamp": "2026-02-15T17:05:19.132917"}
{"step": 37, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.1691796875, "mean_total_reward": 0.954924609375, "std_total_reward": 0.19591740122909576, "total_tokens": 17324, "mean_completion_length": 346.48, "median_completion_length": 335, "min_completion_length": 101, "max_completion_length": 1161, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08076202723527084, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.70287537574768, "avg_step_time_sec": 14.439941779987231, "gpu_allocated_gb": 32.439785472, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 534.5884246826172, "eta_sec": 2353.710510137919, "timestamp": "2026-02-15T17:05:35.653951"}
{"step": 38, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.92, "mean_correctness_reward": 0.92, "mean_length_penalty": -0.15390625, "mean_total_reward": 0.9153828125, "std_total_reward": 0.27124252070825616, "total_tokens": 15760, "mean_completion_length": 315.2, "median_completion_length": 300, "min_completion_length": 173, "max_completion_length": 553, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09856105543473774, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 12.678513765335083, "avg_step_time_sec": 14.393588411180597, "gpu_allocated_gb": 32.438524928, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 547.2754552364349, "eta_sec": 2331.7613226112567, "timestamp": "2026-02-15T17:05:44.959270"}
{"step": 39, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.21919921875, "mean_total_reward": 0.7534240234375, "std_total_reward": 0.4281512548056955, "total_tokens": 22446, "mean_completion_length": 448.92, "median_completion_length": 438, "min_completion_length": 114, "max_completion_length": 1014, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08101502014615693, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 7.138401985168457, "avg_step_time_sec": 14.207557990000797, "gpu_allocated_gb": 32.43772672, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 554.4224197864532, "eta_sec": 2287.4168363901285, "timestamp": "2026-02-15T17:05:55.172641"}
{"step": 40, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.1855078125, "mean_total_reward": 0.754434765625, "std_total_reward": 0.4278676224005486, "total_tokens": 18996, "mean_completion_length": 379.92, "median_completion_length": 381, "min_completion_length": 152, "max_completion_length": 774, "frac_has_think": 1.0, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.12068516983512502, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 11.574692487716675, "avg_step_time_sec": 14.141736352443695, "gpu_allocated_gb": 32.438297088, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 566.0044400691986, "eta_sec": 2262.677816390991, "timestamp": "2026-02-15T17:06:05.236500"}
{"step": 41, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.129716796875, "mean_total_reward": 0.99610849609375, "std_total_reward": 0.002529639872149136, "total_tokens": 13283, "mean_completion_length": 265.66, "median_completion_length": 232, "min_completion_length": 48, "max_completion_length": 691, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.00047882961459537286, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 9.597334861755371, "avg_step_time_sec": 14.030897291695199, "gpu_allocated_gb": 32.438071808, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 575.6095125675201, "eta_sec": 2230.9126693795365, "timestamp": "2026-02-15T17:06:13.845185"}
{"step": 42, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.128935546875, "mean_total_reward": 0.9761319335937501, "std_total_reward": 0.14053573128488853, "total_tokens": 13203, "mean_completion_length": 264.06, "median_completion_length": 292, "min_completion_length": 120, "max_completion_length": 517, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06043519474883794, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.99330472946167, "avg_step_time_sec": 13.887145087832497, "gpu_allocated_gb": 32.437865472, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 583.6113681793213, "eta_sec": 2194.1689238775343, "timestamp": "2026-02-15T17:06:21.164613"}
{"step": 43, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.168232421875, "mean_total_reward": 0.73495302734375, "std_total_reward": 0.43983664049261945, "total_tokens": 17227, "mean_completion_length": 344.54, "median_completion_length": 229, "min_completion_length": 129, "max_completion_length": 787, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09201353275804945, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 6.713858127593994, "avg_step_time_sec": 13.720324460850206, "gpu_allocated_gb": 32.437622784, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 590.3326845169067, "eta_sec": 2154.0909403534824, "timestamp": "2026-02-15T17:06:29.524538"}
{"step": 44, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.179287109375, "mean_total_reward": 0.97462138671875, "std_total_reward": 0.13998922451379872, "total_tokens": 18359, "mean_completion_length": 367.18, "median_completion_length": 317, "min_completion_length": 154, "max_completion_length": 742, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.060329960056638134, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 9.466092824935913, "avg_step_time_sec": 13.62363737821579, "gpu_allocated_gb": 32.438045696, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 599.8064057826996, "eta_sec": 2125.287431001663, "timestamp": "2026-02-15T17:06:38.687351"}
{"step": 45, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.121748046875, "mean_total_reward": 0.95634755859375, "std_total_reward": 0.19608824109243628, "total_tokens": 12467, "mean_completion_length": 249.34, "median_completion_length": 207, "min_completion_length": 105, "max_completion_length": 487, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08044798097179218, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 8.817632913589478, "avg_step_time_sec": 13.516837279001871, "gpu_allocated_gb": 32.4379392, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 608.6330788135529, "eta_sec": 2095.10977824529, "timestamp": "2026-02-15T17:06:45.626988"}
{"step": 46, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.2817578125, "mean_total_reward": 0.591547265625, "std_total_reward": 0.49522139163784656, "total_tokens": 28852, "mean_completion_length": 577.04, "median_completion_length": 277, "min_completion_length": 144, "max_completion_length": 2048, "frac_has_think": 0.9, "frac_has_boxed": 0.9, "mean_reward_std_per_prompt": 0.0013965548972027925, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 6.074015378952026, "avg_step_time_sec": 13.355036802913832, "gpu_allocated_gb": 32.437590528, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 614.7146883010864, "eta_sec": 2056.6756676487303, "timestamp": "2026-02-15T17:07:02.461903"}
{"step": 47, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.155068359375, "mean_total_reward": 0.71534794921875, "std_total_reward": 0.447538970048334, "total_tokens": 15879, "mean_completion_length": 317.58, "median_completion_length": 267, "min_completion_length": 73, "max_completion_length": 758, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.16030440882588412, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.054283380508423, "avg_step_time_sec": 13.561403751373291, "gpu_allocated_gb": 32.439798784, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 637.7776284217834, "eta_sec": 2074.8947739601135, "timestamp": "2026-02-15T17:07:16.565740"}
{"step": 48, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.24724609375, "mean_total_reward": 0.8125826171875, "std_total_reward": 0.38604277460722736, "total_tokens": 25318, "mean_completion_length": 506.36, "median_completion_length": 498, "min_completion_length": 130, "max_completion_length": 1032, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.21224784111930664, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 9.143402099609375, "avg_step_time_sec": 13.469362050294876, "gpu_allocated_gb": 32.437939712, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 646.9295003414154, "eta_sec": 2047.3430316448212, "timestamp": "2026-02-15T17:07:27.970414"}
{"step": 49, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.24638671875, "mean_total_reward": 0.7126083984375, "std_total_reward": 0.45021708422159573, "total_tokens": 25230, "mean_completion_length": 504.6, "median_completion_length": 386, "min_completion_length": 121, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.09927211152992681, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 11.933372259140015, "avg_step_time_sec": 13.438015319863144, "gpu_allocated_gb": 32.438320128, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 658.8708398342133, "eta_sec": 2029.1403132993348, "timestamp": "2026-02-15T17:07:45.776099"}
{"step": 50, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.115908203125, "mean_total_reward": 0.99652275390625, "std_total_reward": 0.001084850622520999, "total_tokens": 11869, "mean_completion_length": 237.38, "median_completion_length": 230, "min_completion_length": 106, "max_completion_length": 468, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0006812975091655951, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 21.947365283966064, "avg_step_time_sec": 13.608202319145203, "gpu_allocated_gb": 32.439745536, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 680.825941324234, "eta_sec": 2041.2303478717804, "timestamp": "2026-02-15T17:07:57.934252"}
{"step": 51, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.277568359375, "mean_total_reward": 0.65167294921875, "std_total_reward": 0.47900504623941537, "total_tokens": 28423, "mean_completion_length": 568.46, "median_completion_length": 314, "min_completion_length": 114, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.09344736836668212, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 5.6472086906433105, "avg_step_time_sec": 13.452104404860853, "gpu_allocated_gb": 32.437524992, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 686.8443427085876, "eta_sec": 2004.363556324267, "timestamp": "2026-02-15T17:08:14.454279"}
{"step": 52, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.86, "mean_correctness_reward": 0.86, "mean_length_penalty": -0.132685546875, "mean_total_reward": 0.85601943359375, "std_total_reward": 0.34719165552337783, "total_tokens": 13587, "mean_completion_length": 271.74, "median_completion_length": 258, "min_completion_length": 131, "max_completion_length": 479, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1583259568505154, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.70257067680359, "avg_step_time_sec": 13.629997987013597, "gpu_allocated_gb": 32.439784448, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 709.5560882091522, "eta_sec": 2017.2397020780124, "timestamp": "2026-02-15T17:08:26.514894"}
{"step": 53, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.44, "mean_correctness_reward": 0.44, "mean_length_penalty": -0.172685546875, "mean_total_reward": 0.43481943359375, "std_total_reward": 0.49638157636050284, "total_tokens": 17683, "mean_completion_length": 353.66, "median_completion_length": 346, "min_completion_length": 142, "max_completion_length": 694, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15201999107427225, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 6.377763748168945, "avg_step_time_sec": 13.49316337873351, "gpu_allocated_gb": 32.437566464, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 715.9419686794281, "eta_sec": 1983.4950166738258, "timestamp": "2026-02-15T17:08:34.463110"}
{"step": 54, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.1730859375, "mean_total_reward": 0.774807421875, "std_total_reward": 0.4156760451880188, "total_tokens": 17724, "mean_completion_length": 354.48, "median_completion_length": 246, "min_completion_length": 106, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06170583148473892, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.254777908325195, "avg_step_time_sec": 13.396156240392614, "gpu_allocated_gb": 32.43785216, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 724.2075290679932, "eta_sec": 1955.8388110973217, "timestamp": "2026-02-15T17:08:50.205469"}
{"step": 55, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.29736328125, "mean_total_reward": 0.7510791015625, "std_total_reward": 0.43400350579260943, "total_tokens": 30450, "mean_completion_length": 609, "median_completion_length": 314, "min_completion_length": 111, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.88, "mean_reward_std_per_prompt": 0.1788288874619271, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 21.01055121421814, "avg_step_time_sec": 13.53459978537126, "gpu_allocated_gb": 32.439744512, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 745.2255473136902, "eta_sec": 1962.5169688788326, "timestamp": "2026-02-15T17:09:13.069489"}
{"step": 56, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.2104296875, "mean_total_reward": 0.993687109375, "std_total_reward": 0.0023247175721937217, "total_tokens": 21548, "mean_completion_length": 430.96, "median_completion_length": 384, "min_completion_length": 232, "max_completion_length": 1166, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0015544794174593246, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.84362816810608, "avg_step_time_sec": 13.700832435062953, "gpu_allocated_gb": 32.439753728, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 768.077930688858, "eta_sec": 1972.9198706490652, "timestamp": "2026-02-15T17:09:29.525860"}
{"step": 57, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.68, "mean_correctness_reward": 0.68, "mean_length_penalty": -0.109111328125, "mean_total_reward": 0.67672666015625, "std_total_reward": 0.4666242699277773, "total_tokens": 11173, "mean_completion_length": 223.46, "median_completion_length": 231, "min_completion_length": 105, "max_completion_length": 415, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.178144156341686, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 12.944120645523071, "avg_step_time_sec": 13.68755678963243, "gpu_allocated_gb": 32.438539264, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 781.0306453704834, "eta_sec": 1957.3206209174375, "timestamp": "2026-02-15T17:09:37.502833"}
{"step": 58, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.124638671875, "mean_total_reward": 0.97626083984375, "std_total_reward": 0.14037031731574895, "total_tokens": 12763, "mean_completion_length": 255.26, "median_completion_length": 257, "min_completion_length": 135, "max_completion_length": 431, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06039362298671814, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 5.8642578125, "avg_step_time_sec": 13.552672324509457, "gpu_allocated_gb": 32.43753216, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 786.9030754566193, "eta_sec": 1924.479470080343, "timestamp": "2026-02-15T17:09:43.510997"}
{"step": 59, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.151689453125, "mean_total_reward": 0.73544931640625, "std_total_reward": 0.43824753713649095, "total_tokens": 15533, "mean_completion_length": 310.66, "median_completion_length": 297, "min_completion_length": 143, "max_completion_length": 524, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1402576584827133, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 5.405923366546631, "avg_step_time_sec": 13.414591833696527, "gpu_allocated_gb": 32.437487104, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 792.3174157142639, "eta_sec": 1891.4574485512103, "timestamp": "2026-02-15T17:09:49.688096"}
{"step": 60, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.2677734375, "mean_total_reward": 0.771966796875, "std_total_reward": 0.4231103006342436, "total_tokens": 27420, "mean_completion_length": 548.4, "median_completion_length": 255, "min_completion_length": 89, "max_completion_length": 2048, "frac_has_think": 0.92, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.06133584598021385, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 6.53477144241333, "avg_step_time_sec": 13.299928160508474, "gpu_allocated_gb": 32.437615104, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 798.8594746589661, "eta_sec": 1861.9899424711864, "timestamp": "2026-02-15T17:10:06.468138"}
{"step": 61, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.298046875, "mean_total_reward": 0.81105859375, "std_total_reward": 0.39330529155517735, "total_tokens": 30520, "mean_completion_length": 610.4, "median_completion_length": 292, "min_completion_length": 144, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.06071621487454857, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.61602234840393, "avg_step_time_sec": 13.452651016047744, "gpu_allocated_gb": 32.439753728, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 821.4850261211395, "eta_sec": 1869.9184912306364, "timestamp": "2026-02-15T17:10:29.328558"}
{"step": 62, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.1509765625, "mean_total_reward": 0.795470703125, "std_total_reward": 0.40159160248741244, "total_tokens": 15460, "mean_completion_length": 309.2, "median_completion_length": 197, "min_completion_length": 84, "max_completion_length": 902, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12022091479391064, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.75522518157959, "avg_step_time_sec": 13.602692534846645, "gpu_allocated_gb": 32.43974144, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 844.249388217926, "eta_sec": 1877.171569808837, "timestamp": "2026-02-15T17:10:43.884964"}
{"step": 63, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.191767578125, "mean_total_reward": 0.97424697265625, "std_total_reward": 0.14046455162246044, "total_tokens": 19637, "mean_completion_length": 392.74, "median_completion_length": 396, "min_completion_length": 204, "max_completion_length": 636, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06057976960830671, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 10.486878395080566, "avg_step_time_sec": 13.55323516754877, "gpu_allocated_gb": 32.438157824, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 854.7446565628052, "eta_sec": 1856.7932179541815, "timestamp": "2026-02-15T17:10:53.171713"}
{"step": 64, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.13681640625, "mean_total_reward": 0.6958955078125, "std_total_reward": 0.45929681034051995, "total_tokens": 14010, "mean_completion_length": 280.2, "median_completion_length": 290, "min_completion_length": 113, "max_completion_length": 482, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1006096940536557, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.159366846084595, "avg_step_time_sec": 13.468955975025892, "gpu_allocated_gb": 32.43782912, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 862.9114878177643, "eta_sec": 1831.7780126035213, "timestamp": "2026-02-15T17:11:00.034136"}
{"step": 65, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.92, "mean_correctness_reward": 0.92, "mean_length_penalty": -0.1672265625, "mean_total_reward": 0.914983203125, "std_total_reward": 0.2709492764513457, "total_tokens": 17124, "mean_completion_length": 342.48, "median_completion_length": 293, "min_completion_length": 173, "max_completion_length": 881, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.200603091462395, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.1556103229522705, "avg_step_time_sec": 13.35644296499399, "gpu_allocated_gb": 32.437550592, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 869.0753223896027, "eta_sec": 1803.1198002741887, "timestamp": "2026-02-15T17:11:08.895458"}
{"step": 66, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.68, "mean_correctness_reward": 0.68, "mean_length_penalty": -0.1420703125, "mean_total_reward": 0.675737890625, "std_total_reward": 0.46718016230879345, "total_tokens": 14548, "mean_completion_length": 290.96, "median_completion_length": 205, "min_completion_length": 50, "max_completion_length": 1050, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.16041915817506086, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 10.371685266494751, "avg_step_time_sec": 13.311219363501579, "gpu_allocated_gb": 32.438168064, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 879.4559032917023, "eta_sec": 1783.7033947092116, "timestamp": "2026-02-15T17:11:19.939291"}
{"step": 67, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.273623046875, "mean_total_reward": 0.83179130859375, "std_total_reward": 0.37614363513574167, "total_tokens": 28019, "mean_completion_length": 560.38, "median_completion_length": 228, "min_completion_length": 70, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.08090975295415724, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 11.485013484954834, "avg_step_time_sec": 13.283962559344163, "gpu_allocated_gb": 32.438359552, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.30938112, "elapsed_sec": 890.9556956291199, "eta_sec": 1766.7670203927737, "timestamp": "2026-02-15T17:11:38.725152"}
{"step": 68, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.114140625, "mean_total_reward": 0.93657578125, "std_total_reward": 0.23704987359517285, "total_tokens": 11688, "mean_completion_length": 233.76, "median_completion_length": 206, "min_completion_length": 80, "max_completion_length": 482, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09221226318090342, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.10072684288025, "avg_step_time_sec": 13.4283267399844, "gpu_allocated_gb": 32.43982848, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 914.064124584198, "eta_sec": 1772.5391296779408, "timestamp": "2026-02-15T17:11:51.350484"}
{"step": 69, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.112197265625, "mean_total_reward": 0.97663408203125, "std_total_reward": 0.14381045652331123, "total_tokens": 11489, "mean_completion_length": 229.78, "median_completion_length": 178, "min_completion_length": 56, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.061971836890480136, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.28431510925293, "avg_step_time_sec": 13.324790339539016, "gpu_allocated_gb": 32.437560832, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 920.3566009998322, "eta_sec": 1745.5475344796112, "timestamp": "2026-02-15T17:12:05.774241"}
{"step": 70, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.193037109375, "mean_total_reward": 0.47420888671875, "std_total_reward": 0.5003824426796382, "total_tokens": 19767, "mean_completion_length": 395.34, "median_completion_length": 342, "min_completion_length": 135, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.17997554951063605, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 20.45881748199463, "avg_step_time_sec": 13.426705013002667, "gpu_allocated_gb": 32.439769088, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 940.8231899738312, "eta_sec": 1745.4716516903468, "timestamp": "2026-02-15T17:12:26.988244"}
{"step": 71, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.52, "mean_correctness_reward": 0.52, "mean_length_penalty": -0.45638671875, "mean_total_reward": 0.5063083984375, "std_total_reward": 0.504012592065616, "total_tokens": 46734, "mean_completion_length": 934.68, "median_completion_length": 646, "min_completion_length": 148, "max_completion_length": 2048, "frac_has_think": 0.82, "frac_has_boxed": 0.82, "mean_reward_std_per_prompt": 0.15333813452322304, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 21.264347553253174, "avg_step_time_sec": 13.537094344555491, "gpu_allocated_gb": 32.43977728, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 962.0971949100494, "eta_sec": 1746.2851704476584, "timestamp": "2026-02-15T17:12:51.344096"}
{"step": 72, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.11966796875, "mean_total_reward": 0.9564099609375, "std_total_reward": 0.19629421171184522, "total_tokens": 12254, "mean_completion_length": 245.08, "median_completion_length": 218, "min_completion_length": 62, "max_completion_length": 493, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08039962210056047, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 24.571683168411255, "avg_step_time_sec": 13.6903525226646, "gpu_allocated_gb": 32.439804928, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 986.6774854660034, "eta_sec": 1752.3651229010688, "timestamp": "2026-02-15T17:13:03.665504"}
{"step": 73, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.106533203125, "mean_total_reward": 0.99680400390625, "std_total_reward": 0.0015257138848265889, "total_tokens": 10909, "mean_completion_length": 218.18, "median_completion_length": 198, "min_completion_length": 81, "max_completion_length": 570, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0007561426436146918, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.41499924659729, "avg_step_time_sec": 13.590690149019842, "gpu_allocated_gb": 32.437581312, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 993.1029984951019, "eta_sec": 1726.01764892552, "timestamp": "2026-02-15T17:13:10.356085"}
{"step": 74, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.44, "mean_correctness_reward": 0.44, "mean_length_penalty": -0.316630859375, "mean_total_reward": 0.43050107421875, "std_total_reward": 0.5008658963853779, "total_tokens": 32423, "mean_completion_length": 648.46, "median_completion_length": 302, "min_completion_length": 107, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.15369543362676688, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 6.4624738693237305, "avg_step_time_sec": 13.494362901996922, "gpu_allocated_gb": 32.437666304, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 999.573323726654, "eta_sec": 1700.2897256516121, "timestamp": "2026-02-15T17:13:27.567904"}
{"step": 75, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.140576171875, "mean_total_reward": 0.61578271484375, "std_total_reward": 0.4865586097667118, "total_tokens": 14395, "mean_completion_length": 287.9, "median_completion_length": 259, "min_completion_length": 73, "max_completion_length": 965, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.21171594986385203, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.264724731445312, "avg_step_time_sec": 13.624634393056233, "gpu_allocated_gb": 32.43980288, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1022.8470118045807, "eta_sec": 1703.079299132029, "timestamp": "2026-02-15T17:13:42.597872"}
{"step": 76, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.4083984375, "mean_total_reward": 0.687748046875, "std_total_reward": 0.46334087521134676, "total_tokens": 41820, "mean_completion_length": 836.4, "median_completion_length": 868, "min_completion_length": 104, "max_completion_length": 2048, "frac_has_think": 0.94, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.23884528113637257, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 10.564726829528809, "avg_step_time_sec": 13.584372451430873, "gpu_allocated_gb": 32.438272512, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1033.4197688102722, "eta_sec": 1684.4621839774284, "timestamp": "2026-02-15T17:14:02.120463"}
{"step": 77, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.096171875, "mean_total_reward": 0.75711484375, "std_total_reward": 0.4273915162618996, "total_tokens": 9848, "mean_completion_length": 196.96, "median_completion_length": 200, "min_completion_length": 39, "max_completion_length": 423, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.17843079559659256, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.85573649406433, "avg_step_time_sec": 13.717766789646891, "gpu_allocated_gb": 32.439794688, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1057.283627986908, "eta_sec": 1687.2853151265676, "timestamp": "2026-02-15T17:14:13.689570"}
{"step": 78, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.86, "mean_correctness_reward": 0.86, "mean_length_penalty": -0.231123046875, "mean_total_reward": 0.85306630859375, "std_total_reward": 0.3486207167231201, "total_tokens": 23667, "mean_completion_length": 473.34, "median_completion_length": 206, "min_completion_length": 39, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.23902148708046628, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 5.656531095504761, "avg_step_time_sec": 13.614417614080967, "gpu_allocated_gb": 32.437491712, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1062.9489448070526, "eta_sec": 1660.958948917878, "timestamp": "2026-02-15T17:14:30.087703"}
{"step": 79, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.097490234375, "mean_total_reward": 0.77707529296875, "std_total_reward": 0.4157384396807742, "total_tokens": 9983, "mean_completion_length": 199.66, "median_completion_length": 129, "min_completion_length": 73, "max_completion_length": 631, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06037498717489548, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.190916538238525, "avg_step_time_sec": 13.72298089160195, "gpu_allocated_gb": 32.439788544, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1085.1481297016144, "eta_sec": 1660.4806878838358, "timestamp": "2026-02-15T17:14:42.746779"}
{"step": 80, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.21712890625, "mean_total_reward": 0.5534861328125, "std_total_reward": 0.5000376710957659, "total_tokens": 22234, "mean_completion_length": 444.68, "median_completion_length": 196, "min_completion_length": 57, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.23207621332324405, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 7.231448650360107, "avg_step_time_sec": 13.641836738586425, "gpu_allocated_gb": 32.437801472, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1092.3871870040894, "eta_sec": 1637.020408630371, "timestamp": "2026-02-15T17:14:59.469900"}
{"step": 81, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.089912109375, "mean_total_reward": 0.77730263671875, "std_total_reward": 0.4159910166337731, "total_tokens": 9207, "mean_completion_length": 184.14, "median_completion_length": 131, "min_completion_length": 43, "max_completion_length": 594, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06058021813114962, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.201549768447876, "avg_step_time_sec": 13.747512208090889, "gpu_allocated_gb": 32.439779328, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1114.5959768295288, "eta_sec": 1635.9539527628158, "timestamp": "2026-02-15T17:15:11.988923"}
{"step": 82, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.162080078125, "mean_total_reward": 0.95513759765625, "std_total_reward": 0.19555351581385386, "total_tokens": 16597, "mean_completion_length": 331.94, "median_completion_length": 314, "min_completion_length": 130, "max_completion_length": 738, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0806761363627725, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.819223642349243, "avg_step_time_sec": 13.66302088411843, "gpu_allocated_gb": 32.43774208, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1121.4234466552734, "eta_sec": 1612.2364643259748, "timestamp": "2026-02-15T17:15:20.104459"}
{"step": 83, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.2585546875, "mean_total_reward": 0.592243359375, "std_total_reward": 0.49637907239516715, "total_tokens": 26476, "mean_completion_length": 529.52, "median_completion_length": 171, "min_completion_length": 48, "max_completion_length": 2048, "frac_has_think": 0.82, "frac_has_boxed": 0.82, "mean_reward_std_per_prompt": 0.0002471405176552444, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 8.513742208480835, "avg_step_time_sec": 13.600981382002313, "gpu_allocated_gb": 32.437907968, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1129.9450047016144, "eta_sec": 1591.3148216942707, "timestamp": "2026-02-15T17:15:37.877639"}
{"step": 84, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.121572265625, "mean_total_reward": 0.75635283203125, "std_total_reward": 0.4271654714728982, "total_tokens": 12449, "mean_completion_length": 248.98, "median_completion_length": 225, "min_completion_length": 57, "max_completion_length": 1450, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0812432488685652, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.981243133544922, "avg_step_time_sec": 13.712651164758773, "gpu_allocated_gb": 32.439773184, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1152.9345467090607, "eta_sec": 1590.6675351120177, "timestamp": "2026-02-15T17:15:55.191921"}
{"step": 85, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.64, "mean_correctness_reward": 0.64, "mean_length_penalty": -0.113154296875, "mean_total_reward": 0.63660537109375, "std_total_reward": 0.4797368087409031, "total_tokens": 11587, "mean_completion_length": 231.74, "median_completion_length": 226, "min_completion_length": 81, "max_completion_length": 495, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.21835882916832236, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 15.02047085762024, "avg_step_time_sec": 13.728037278792437, "gpu_allocated_gb": 32.438951424, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1167.9632744789124, "eta_sec": 1578.7242870611303, "timestamp": "2026-02-15T17:16:04.893889"}
{"step": 86, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.165615234375, "mean_total_reward": 0.95503154296875, "std_total_reward": 0.1975587537440121, "total_tokens": 16959, "mean_completion_length": 339.18, "median_completion_length": 231, "min_completion_length": 70, "max_completion_length": 1513, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12125528642891073, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 5.996105194091797, "avg_step_time_sec": 13.638131091761034, "gpu_allocated_gb": 32.437574656, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1173.967216014862, "eta_sec": 1554.746944460758, "timestamp": "2026-02-15T17:16:17.111425"}
{"step": 87, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.64, "mean_correctness_reward": 0.64, "mean_length_penalty": -0.23890625, "mean_total_reward": 0.6328328125, "std_total_reward": 0.48419718561630853, "total_tokens": 24464, "mean_completion_length": 489.28, "median_completion_length": 300, "min_completion_length": 71, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.08068851805901632, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 15.994692325592041, "avg_step_time_sec": 13.665218002494724, "gpu_allocated_gb": 32.439030784, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.322177536, "elapsed_sec": 1189.9695336818695, "eta_sec": 1544.1696342819039, "timestamp": "2026-02-15T17:16:37.278908"}
{"step": 88, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.173037109375, "mean_total_reward": 0.79480888671875, "std_total_reward": 0.4021377914867926, "total_tokens": 17719, "mean_completion_length": 354.38, "median_completion_length": 265, "min_completion_length": 76, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.2320808860116524, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.77232027053833, "avg_step_time_sec": 13.76870780099522, "gpu_allocated_gb": 32.43982848, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1212.7503316402435, "eta_sec": 1542.0952737114646, "timestamp": "2026-02-15T17:16:59.026739"}
{"step": 89, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.10533203125, "mean_total_reward": 0.7768400390625, "std_total_reward": 0.4142093933250055, "total_tokens": 10786, "mean_completion_length": 215.72, "median_completion_length": 206, "min_completion_length": 104, "max_completion_length": 670, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06067016835763728, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 21.6082866191864, "avg_step_time_sec": 13.856792956255795, "gpu_allocated_gb": 32.439794688, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1234.3677797317505, "eta_sec": 1538.1040181443932, "timestamp": "2026-02-15T17:17:12.096717"}
{"step": 90, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.0821484375, "mean_total_reward": 0.977535546875, "std_total_reward": 0.14026916179225146, "total_tokens": 8412, "mean_completion_length": 168.24, "median_completion_length": 175, "min_completion_length": 82, "max_completion_length": 296, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0605470166982676, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.335855960845947, "avg_step_time_sec": 13.78433810075124, "gpu_allocated_gb": 32.437809152, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1241.7119114398956, "eta_sec": 1516.2771910826364, "timestamp": "2026-02-15T17:17:17.238431"}
{"step": 91, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.36, "mean_correctness_reward": 0.36, "mean_length_penalty": -0.142822265625, "mean_total_reward": 0.35571533203125, "std_total_reward": 0.48027204391395206, "total_tokens": 14625, "mean_completion_length": 292.5, "median_completion_length": 276, "min_completion_length": 104, "max_completion_length": 604, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08089387510612192, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 3.843029022216797, "avg_step_time_sec": 13.67509294604207, "gpu_allocated_gb": 32.437284864, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1245.5628044605255, "eta_sec": 1490.5851311185857, "timestamp": "2026-02-15T17:17:23.405850"}
{"step": 92, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.083994140625, "mean_total_reward": 0.83748017578125, "std_total_reward": 0.3667638303277478, "total_tokens": 8601, "mean_completion_length": 172.02, "median_completion_length": 158, "min_completion_length": 100, "max_completion_length": 319, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.19195533194358114, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.483137369155884, "avg_step_time_sec": 13.607789081075918, "gpu_allocated_gb": 32.437798912, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1253.0535833835602, "eta_sec": 1469.6412207561991, "timestamp": "2026-02-15T17:17:28.767748"}
{"step": 93, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.28, "mean_correctness_reward": 0.28, "mean_length_penalty": -0.184365234375, "mean_total_reward": 0.27446904296875, "std_total_reward": 0.44947729871474923, "total_tokens": 18879, "mean_completion_length": 377.58, "median_completion_length": 277, "min_completion_length": 111, "max_completion_length": 1327, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09847074711039865, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 4.264122247695923, "avg_step_time_sec": 13.507319545233122, "gpu_allocated_gb": 32.437362176, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1257.327040195465, "eta_sec": 1445.283191339944, "timestamp": "2026-02-15T17:17:39.613675"}
{"step": 94, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.251552734375, "mean_total_reward": 0.79245341796875, "std_total_reward": 0.4095108595599492, "total_tokens": 25759, "mean_completion_length": 515.18, "median_completion_length": 190, "min_completion_length": 71, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.0017301557468143748, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 14.56169319152832, "avg_step_time_sec": 13.518536286151155, "gpu_allocated_gb": 32.43878656, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1271.8990197181702, "eta_sec": 1432.9648463320225, "timestamp": "2026-02-15T17:17:59.344920"}
{"step": 95, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.125673828125, "mean_total_reward": 0.7562297851562501, "std_total_reward": 0.42607792220114926, "total_tokens": 12869, "mean_completion_length": 257.38, "median_completion_length": 172, "min_completion_length": 53, "max_completion_length": 693, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08049673746731752, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 23.02373957633972, "avg_step_time_sec": 13.618591057626825, "gpu_allocated_gb": 32.43981312, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1294.9311397075653, "eta_sec": 1429.9520610508166, "timestamp": "2026-02-15T17:18:13.336394"}
{"step": 96, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.64, "mean_correctness_reward": 0.64, "mean_length_penalty": -0.259013671875, "mean_total_reward": 0.63222958984375, "std_total_reward": 0.4831685830910891, "total_tokens": 26523, "mean_completion_length": 530.46, "median_completion_length": 232, "min_completion_length": 111, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.17886958006302836, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.619625568389893, "avg_step_time_sec": 13.566518500447273, "gpu_allocated_gb": 32.437991936, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1303.559157371521, "eta_sec": 1410.9179240465164, "timestamp": "2026-02-15T17:18:31.102705"}
{"step": 97, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.19533203125, "mean_total_reward": 0.8341400390625, "std_total_reward": 0.3700311219408902, "total_tokens": 20002, "mean_completion_length": 400.04, "median_completion_length": 247, "min_completion_length": 73, "max_completion_length": 1552, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15163305107416436, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.64530849456787, "avg_step_time_sec": 13.660114273582536, "gpu_allocated_gb": 32.439784448, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1326.212647676468, "eta_sec": 1406.9917701790012, "timestamp": "2026-02-15T17:18:50.233979"}
{"step": 98, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.1629296875, "mean_total_reward": 0.495112109375, "std_total_reward": 0.5011819100705873, "total_tokens": 16684, "mean_completion_length": 333.68, "median_completion_length": 293, "min_completion_length": 134, "max_completion_length": 741, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2604731271643486, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 16.985448837280273, "avg_step_time_sec": 13.694046258926392, "gpu_allocated_gb": 32.439080448, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1343.2063863277435, "eta_sec": 1396.792718410492, "timestamp": "2026-02-15T17:19:01.973306"}
{"step": 99, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.078759765625, "mean_total_reward": 0.7776372070312499, "std_total_reward": 0.4145597174726641, "total_tokens": 8065, "mean_completion_length": 161.3, "median_completion_length": 148, "min_completion_length": 85, "max_completion_length": 389, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.060419872255955534, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.808224439620972, "avg_step_time_sec": 13.644694523377852, "gpu_allocated_gb": 32.437947392, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1352.0231387615204, "eta_sec": 1378.1141468611631, "timestamp": "2026-02-15T17:19:08.191945"}
{"step": 100, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.154462890625, "mean_total_reward": 0.61536611328125, "std_total_reward": 0.4877802661815104, "total_tokens": 15817, "mean_completion_length": 316.34, "median_completion_length": 165, "min_completion_length": 66, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.06108048241688589, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 4.905955076217651, "avg_step_time_sec": 13.55730712890625, "gpu_allocated_gb": 32.437446656, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1356.940489768982, "eta_sec": 1355.730712890625, "timestamp": "2026-02-15T17:19:23.723128"}
{"step": 101, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.10318359375, "mean_total_reward": 0.9569044921875, "std_total_reward": 0.1962082050500584, "total_tokens": 10566, "mean_completion_length": 211.32, "median_completion_length": 203, "min_completion_length": 70, "max_completion_length": 430, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12019038637585708, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 21.843125820159912, "avg_step_time_sec": 13.639344937730543, "gpu_allocated_gb": 32.439755776, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1379.1148586273193, "eta_sec": 1350.2951488353237, "timestamp": "2026-02-15T17:19:35.611285"}
{"step": 102, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.1276171875, "mean_total_reward": 0.996171484375, "std_total_reward": 0.0015141688448285426, "total_tokens": 13068, "mean_completion_length": 261.36, "median_completion_length": 223, "min_completion_length": 136, "max_completion_length": 552, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0006116273346911188, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 5.278742551803589, "avg_step_time_sec": 13.557378247672437, "gpu_allocated_gb": 32.4374784, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1384.4031398296356, "eta_sec": 1328.6230682718988, "timestamp": "2026-02-15T17:19:41.818248"}
{"step": 103, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.64, "mean_correctness_reward": 0.64, "mean_length_penalty": -0.249541015625, "mean_total_reward": 0.63251376953125, "std_total_reward": 0.48361604427751653, "total_tokens": 25553, "mean_completion_length": 511.06, "median_completion_length": 371, "min_completion_length": 141, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.15252799034569905, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 6.633204460144043, "avg_step_time_sec": 13.490153259444005, "gpu_allocated_gb": 32.437654528, "gpu_reserved_gb": 81.176559616, "gpu_peak_gb": 74.324004864, "elapsed_sec": 1391.044228553772, "eta_sec": 1308.5448661660685, "timestamp": "2026-02-15T17:19:58.635564"}
{"step": 104, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.44, "mean_correctness_reward": 0.44, "mean_length_penalty": -0.2382421875, "mean_total_reward": 0.432852734375, "std_total_reward": 0.4988684877718133, "total_tokens": 24396, "mean_completion_length": 487.92, "median_completion_length": 299, "min_completion_length": 96, "max_completion_length": 2048, "frac_has_think": 0.96, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.19795319526567706, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 23.002430200576782, "avg_step_time_sec": 13.581617460801052, "gpu_allocated_gb": 32.439831552, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1414.0550000667572, "eta_sec": 1303.8352762369009, "timestamp": "2026-02-15T17:20:21.297872"}
{"step": 105, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.13400390625, "mean_total_reward": 0.8359798828125, "std_total_reward": 0.36641940217568547, "total_tokens": 13722, "mean_completion_length": 274.44, "median_completion_length": 261, "min_completion_length": 137, "max_completion_length": 532, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.24014216065700736, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.237666130065918, "avg_step_time_sec": 13.664056019555956, "gpu_allocated_gb": 32.439747584, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1436.3023352622986, "eta_sec": 1298.0853218578159, "timestamp": "2026-02-15T17:20:33.674190"}
{"step": 106, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.18908203125, "mean_total_reward": 0.8143275390625, "std_total_reward": 0.3869956374649962, "total_tokens": 19362, "mean_completion_length": 387.24, "median_completion_length": 243, "min_completion_length": 101, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.14072845826672234, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.6972761154174805, "avg_step_time_sec": 13.598331680837667, "gpu_allocated_gb": 32.43767296, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1443.010794878006, "eta_sec": 1278.2431779987407, "timestamp": "2026-02-15T17:20:49.931197"}
{"step": 107, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.86, "mean_correctness_reward": 0.86, "mean_length_penalty": -0.11248046875, "mean_total_reward": 0.8566255859375, "std_total_reward": 0.34693967575718687, "total_tokens": 11518, "mean_completion_length": 230.36, "median_completion_length": 206, "min_completion_length": 77, "max_completion_length": 860, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.18010620640116926, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 21.93757963180542, "avg_step_time_sec": 13.676268577575684, "gpu_allocated_gb": 32.43976192, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1464.9565455913544, "eta_sec": 1271.8929777145386, "timestamp": "2026-02-15T17:21:03.919001"}
{"step": 108, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.1107421875, "mean_total_reward": 0.556677734375, "std_total_reward": 0.4962188030593395, "total_tokens": 11340, "mean_completion_length": 226.8, "median_completion_length": 223, "min_completion_length": 130, "max_completion_length": 346, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08033319921978943, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 9.311095714569092, "avg_step_time_sec": 13.635850310325623, "gpu_allocated_gb": 32.43810304, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1474.2760426998138, "eta_sec": 1254.4982285499573, "timestamp": "2026-02-15T17:21:10.356654"}
{"step": 109, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.086923828125, "mean_total_reward": 0.99739228515625, "std_total_reward": 0.0012163478594750888, "total_tokens": 8901, "mean_completion_length": 178.02, "median_completion_length": 147, "min_completion_length": 73, "max_completion_length": 407, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.00042861528553225694, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.701759576797485, "avg_step_time_sec": 13.55388617515564, "gpu_allocated_gb": 32.437399552, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1478.9857354164124, "eta_sec": 1233.4036419391632, "timestamp": "2026-02-15T17:21:15.239832"}
{"step": 110, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.18552734375, "mean_total_reward": 0.8144341796875, "std_total_reward": 0.3861694019108199, "total_tokens": 18998, "mean_completion_length": 379.96, "median_completion_length": 299, "min_completion_length": 43, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.27011756406095133, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 5.023463249206543, "avg_step_time_sec": 13.47633687582883, "gpu_allocated_gb": 32.437468672, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1484.0177791118622, "eta_sec": 1212.8703188245947, "timestamp": "2026-02-15T17:21:30.486137"}
{"step": 111, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.15064453125, "mean_total_reward": 0.5954806640625, "std_total_reward": 0.4906295610992488, "total_tokens": 15426, "mean_completion_length": 308.52, "median_completion_length": 198, "min_completion_length": 109, "max_completion_length": 1024, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.19617610193726737, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 21.739143133163452, "avg_step_time_sec": 13.550776571840853, "gpu_allocated_gb": 32.439800832, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1505.7642619609833, "eta_sec": 1206.0191148938359, "timestamp": "2026-02-15T17:21:46.125857"}
{"step": 112, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.14482421875, "mean_total_reward": 0.7356552734375, "std_total_reward": 0.4384350738001476, "total_tokens": 14830, "mean_completion_length": 296.6, "median_completion_length": 258, "min_completion_length": 162, "max_completion_length": 699, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09226120620707524, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 11.291653633117676, "avg_step_time_sec": 13.530605831316539, "gpu_allocated_gb": 32.438330368, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1517.064734697342, "eta_sec": 1190.6933131558553, "timestamp": "2026-02-15T17:21:55.783151"}
{"step": 113, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.36, "mean_correctness_reward": 0.36, "mean_length_penalty": -0.314423828125, "mean_total_reward": 0.35056728515625, "std_total_reward": 0.48342528753920627, "total_tokens": 32197, "mean_completion_length": 643.94, "median_completion_length": 307, "min_completion_length": 155, "max_completion_length": 2048, "frac_has_think": 0.82, "frac_has_boxed": 0.82, "mean_reward_std_per_prompt": 0.19220929950888474, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 8.512138605117798, "avg_step_time_sec": 13.4861946169254, "gpu_allocated_gb": 32.43787264, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1525.5848224163055, "eta_sec": 1173.29893167251, "timestamp": "2026-02-15T17:22:13.719803"}
{"step": 114, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.086953125, "mean_total_reward": 0.97739140625, "std_total_reward": 0.1413386480899994, "total_tokens": 8904, "mean_completion_length": 178.08, "median_completion_length": 174, "min_completion_length": 47, "max_completion_length": 817, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.060787186867221364, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.509433269500732, "avg_step_time_sec": 13.574117763000622, "gpu_allocated_gb": 32.439830528, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1549.1021800041199, "eta_sec": 1167.3741276180535, "timestamp": "2026-02-15T17:22:27.461011"}
{"step": 115, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.163095703125, "mean_total_reward": 0.59510712890625, "std_total_reward": 0.49350701204631614, "total_tokens": 16701, "mean_completion_length": 334.02, "median_completion_length": 140, "min_completion_length": 77, "max_completion_length": 1737, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.001333843983441882, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 8.452432870864868, "avg_step_time_sec": 13.529581372634224, "gpu_allocated_gb": 32.438024192, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1557.5631275177002, "eta_sec": 1150.014416673909, "timestamp": "2026-02-15T17:22:42.284907"}
{"step": 116, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.123896484375, "mean_total_reward": 0.71628310546875, "std_total_reward": 0.44874726707273427, "total_tokens": 12687, "mean_completion_length": 253.74, "median_completion_length": 224, "min_completion_length": 107, "max_completion_length": 857, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.20007619772425375, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 18.63129496574402, "avg_step_time_sec": 13.573561662229999, "gpu_allocated_gb": 32.439369216, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1576.2018291950226, "eta_sec": 1140.1791796273199, "timestamp": "2026-02-15T17:22:55.259869"}
{"step": 117, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.25212890625, "mean_total_reward": 0.4924361328125, "std_total_reward": 0.5043883012758105, "total_tokens": 25818, "mean_completion_length": 516.36, "median_completion_length": 249, "min_completion_length": 87, "max_completion_length": 2048, "frac_has_think": 0.88, "frac_has_boxed": 0.88, "mean_reward_std_per_prompt": 0.22160254191996603, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 9.313377618789673, "avg_step_time_sec": 13.537149832798885, "gpu_allocated_gb": 32.438071808, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1585.523520231247, "eta_sec": 1123.5834361223074, "timestamp": "2026-02-15T17:23:12.981376"}
{"step": 118, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.07375, "mean_total_reward": 0.9377875, "std_total_reward": 0.23765678433317952, "total_tokens": 7552, "mean_completion_length": 151.04, "median_completion_length": 135, "min_completion_length": 68, "max_completion_length": 547, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1401871128928594, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 23.10079836845398, "avg_step_time_sec": 13.618197701745114, "gpu_allocated_gb": 32.43980288, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1608.6319434642792, "eta_sec": 1116.6922115430993, "timestamp": "2026-02-15T17:23:25.449843"}
{"step": 119, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.38, "mean_correctness_reward": 0.38, "mean_length_penalty": -0.275224609375, "mean_total_reward": 0.37174326171875, "std_total_reward": 0.4896982556800729, "total_tokens": 28183, "mean_completion_length": 563.66, "median_completion_length": 243, "min_completion_length": 58, "max_completion_length": 2048, "frac_has_think": 0.82, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.06086530011109824, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 5.9382970333099365, "avg_step_time_sec": 13.553660721338096, "gpu_allocated_gb": 32.437627904, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1614.5785384178162, "eta_sec": 1097.8465184283857, "timestamp": "2026-02-15T17:23:42.297779"}
{"step": 120, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.72, "mean_correctness_reward": 0.72, "mean_length_penalty": -0.11126953125, "mean_total_reward": 0.7166619140625, "std_total_reward": 0.44891068422337294, "total_tokens": 11394, "mean_completion_length": 227.88, "median_completion_length": 200, "min_completion_length": 109, "max_completion_length": 765, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.20037895481443405, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.824921369552612, "avg_step_time_sec": 13.630921226739883, "gpu_allocated_gb": 32.43977216, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1637.4095396995544, "eta_sec": 1090.4736981391907, "timestamp": "2026-02-15T17:23:55.811015"}
{"step": 121, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.077197265625, "mean_total_reward": 0.93768408203125, "std_total_reward": 0.23747654971398077, "total_tokens": 7905, "mean_completion_length": 158.1, "median_completion_length": 136, "min_completion_length": 49, "max_completion_length": 318, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.18005514848628978, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 8.343259334564209, "avg_step_time_sec": 13.58722154168058, "gpu_allocated_gb": 32.437954048, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1645.7608923912048, "eta_sec": 1073.390501792766, "timestamp": "2026-02-15T17:24:01.387366"}
{"step": 122, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.12921875, "mean_total_reward": 0.7561234375, "std_total_reward": 0.4285453685112335, "total_tokens": 13232, "mean_completion_length": 264.64, "median_completion_length": 247, "min_completion_length": 59, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.0818475900126249, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 3.998579263687134, "avg_step_time_sec": 13.508626113172438, "gpu_allocated_gb": 32.437313024, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1649.7675116062164, "eta_sec": 1053.6728368274503, "timestamp": "2026-02-15T17:24:15.437168"}
{"step": 123, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.130029296875, "mean_total_reward": 0.65609912109375, "std_total_reward": 0.47499420315693736, "total_tokens": 13315, "mean_completion_length": 266.3, "median_completion_length": 228, "min_completion_length": 85, "max_completion_length": 686, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15848054337808096, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 20.782487869262695, "avg_step_time_sec": 13.567763200620325, "gpu_allocated_gb": 32.439793664, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1670.5577580928802, "eta_sec": 1044.717766447765, "timestamp": "2026-02-15T17:24:28.734438"}
{"step": 124, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.11294921875, "mean_total_reward": 0.7966115234375, "std_total_reward": 0.40185587054885746, "total_tokens": 11566, "mean_completion_length": 231.32, "median_completion_length": 190, "min_completion_length": 84, "max_completion_length": 786, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0007352946594892136, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 7.823310136795044, "avg_step_time_sec": 13.521436966234639, "gpu_allocated_gb": 32.437839872, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1678.3899421691895, "eta_sec": 1027.6292094338326, "timestamp": "2026-02-15T17:24:37.329018"}
{"step": 125, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.086962890625, "mean_total_reward": 0.59739111328125, "std_total_reward": 0.49023599452628913, "total_tokens": 8905, "mean_completion_length": 178.1, "median_completion_length": 187, "min_completion_length": 49, "max_completion_length": 320, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.00037645478488045053, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 8.989054918289185, "avg_step_time_sec": 13.485177909851075, "gpu_allocated_gb": 32.437997568, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1687.3864576816559, "eta_sec": 1011.3883432388307, "timestamp": "2026-02-15T17:24:43.093790"}
{"step": 126, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.19685546875, "mean_total_reward": 0.6140943359375, "std_total_reward": 0.48772756903226605, "total_tokens": 20158, "mean_completion_length": 403.16, "median_completion_length": 229, "min_completion_length": 141, "max_completion_length": 1759, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.14063177377013097, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 4.253022909164429, "avg_step_time_sec": 13.411906838417053, "gpu_allocated_gb": 32.43732736, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1691.6471183300018, "eta_sec": 992.4811060428619, "timestamp": "2026-02-15T17:24:56.629816"}
{"step": 127, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.82, "mean_correctness_reward": 0.82, "mean_length_penalty": -0.07630859375, "mean_total_reward": 0.8177107421875, "std_total_reward": 0.3842397175424599, "total_tokens": 7814, "mean_completion_length": 156.28, "median_completion_length": 162, "min_completion_length": 67, "max_completion_length": 264, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.14016410309292612, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 18.796178102493286, "avg_step_time_sec": 13.454302675142063, "gpu_allocated_gb": 32.439363584, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1710.4511444568634, "eta_sec": 982.1640952853706, "timestamp": "2026-02-15T17:25:05.825462"}
{"step": 128, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.0947265625, "mean_total_reward": 0.397158203125, "std_total_reward": 0.48990748109472415, "total_tokens": 9700, "mean_completion_length": 194, "median_completion_length": 141, "min_completion_length": 80, "max_completion_length": 533, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.19608482345833653, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 3.749967336654663, "avg_step_time_sec": 13.37848755531013, "gpu_allocated_gb": 32.437292032, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1714.2090935707092, "eta_sec": 963.2511039823294, "timestamp": "2026-02-15T17:25:11.282239"}
{"step": 129, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.9, "mean_correctness_reward": 0.9, "mean_length_penalty": -0.084140625, "mean_total_reward": 0.89747578125, "std_total_reward": 0.3009834117737706, "total_tokens": 8616, "mean_completion_length": 172.32, "median_completion_length": 134, "min_completion_length": 57, "max_completion_length": 862, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15811658279140076, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.356400728225708, "avg_step_time_sec": 13.324052773704825, "gpu_allocated_gb": 32.43766528, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1720.5732769966125, "eta_sec": 946.0077469330425, "timestamp": "2026-02-15T17:25:19.337502"}
{"step": 130, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.255263671875, "mean_total_reward": 0.77234208984375, "std_total_reward": 0.42353538154114123, "total_tokens": 26139, "mean_completion_length": 522.78, "median_completion_length": 164, "min_completion_length": 81, "max_completion_length": 2048, "frac_has_think": 0.84, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.1408607598654514, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 9.091346263885498, "avg_step_time_sec": 13.291493492860061, "gpu_allocated_gb": 32.438098432, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1729.6725237369537, "eta_sec": 930.4045445002043, "timestamp": "2026-02-15T17:25:37.608108"}
{"step": 131, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.08111328125, "mean_total_reward": 0.7975666015625, "std_total_reward": 0.4003911311172011, "total_tokens": 8306, "mean_completion_length": 166.12, "median_completion_length": 160, "min_completion_length": 115, "max_completion_length": 243, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.00028698396243721335, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.94737935066223, "avg_step_time_sec": 13.365202545209696, "gpu_allocated_gb": 32.439763968, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1752.6280725002289, "eta_sec": 922.198975619469, "timestamp": "2026-02-15T17:25:47.855797"}
{"step": 132, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.097099609375, "mean_total_reward": 0.93708701171875, "std_total_reward": 0.2381940337312502, "total_tokens": 9943, "mean_completion_length": 198.86, "median_completion_length": 169, "min_completion_length": 94, "max_completion_length": 814, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.14058234583397464, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 3.4233310222625732, "avg_step_time_sec": 13.28988533670252, "gpu_allocated_gb": 32.43722496, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1756.0592379570007, "eta_sec": 903.7122028957714, "timestamp": "2026-02-15T17:25:54.596716"}
{"step": 133, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.11322265625, "mean_total_reward": 0.6166033203125, "std_total_reward": 0.4848427132487644, "total_tokens": 11594, "mean_completion_length": 231.88, "median_completion_length": 213, "min_completion_length": 105, "max_completion_length": 591, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2117002081808476, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 8.774773120880127, "avg_step_time_sec": 13.25593712455348, "gpu_allocated_gb": 32.438048768, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1764.8414611816406, "eta_sec": 888.1477873450832, "timestamp": "2026-02-15T17:26:02.312387"}
{"step": 134, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.175732421875, "mean_total_reward": 0.75472802734375, "std_total_reward": 0.43262080441032874, "total_tokens": 17995, "mean_completion_length": 359.9, "median_completion_length": 140, "min_completion_length": 76, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.12183198390961562, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 6.821574687957764, "avg_step_time_sec": 13.207919494429632, "gpu_allocated_gb": 32.43771136, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1771.6712710857391, "eta_sec": 871.7226866323557, "timestamp": "2026-02-15T17:26:18.376541"}
{"step": 135, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.077734375, "mean_total_reward": 0.65766796875, "std_total_reward": 0.4738506309524398, "total_tokens": 7960, "mean_completion_length": 159.2, "median_completion_length": 143, "min_completion_length": 82, "max_completion_length": 288, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09188246716429604, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 21.73750066757202, "avg_step_time_sec": 13.27110157719365, "gpu_allocated_gb": 32.439754752, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1793.4164814949036, "eta_sec": 862.6216025175872, "timestamp": "2026-02-15T17:26:28.882306"}
{"step": 136, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.099375, "mean_total_reward": 0.59701875, "std_total_reward": 0.4906981186487667, "total_tokens": 10176, "mean_completion_length": 203.52, "median_completion_length": 182, "min_completion_length": 49, "max_completion_length": 1179, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0010085521275192918, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 3.832218885421753, "avg_step_time_sec": 13.201698027989444, "gpu_allocated_gb": 32.437292544, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1797.255258321762, "eta_sec": 844.9086737913244, "timestamp": "2026-02-15T17:26:37.806454"}
{"step": 137, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.56, "mean_correctness_reward": 0.56, "mean_length_penalty": -0.088974609375, "mean_total_reward": 0.55733076171875, "std_total_reward": 0.49689039531883467, "total_tokens": 9111, "mean_completion_length": 182.22, "median_completion_length": 174, "min_completion_length": 56, "max_completion_length": 595, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08050747783671314, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 12.189400911331177, "avg_step_time_sec": 13.194308997940842, "gpu_allocated_gb": 32.438563328, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1809.4538974761963, "eta_sec": 831.241466870273, "timestamp": "2026-02-15T17:26:47.097423"}
{"step": 138, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.117548828125, "mean_total_reward": 0.83647353515625, "std_total_reward": 0.3678304151779677, "total_tokens": 12037, "mean_completion_length": 240.74, "median_completion_length": 147, "min_completion_length": 65, "max_completion_length": 907, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08027682964124287, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.023656368255615, "avg_step_time_sec": 13.149594123812689, "gpu_allocated_gb": 32.437704192, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1816.4852788448334, "eta_sec": 815.2748356763867, "timestamp": "2026-02-15T17:26:55.860096"}
{"step": 139, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.08427734375, "mean_total_reward": 0.9774716796875, "std_total_reward": 0.13980226612109906, "total_tokens": 8630, "mean_completion_length": 172.6, "median_completion_length": 189, "min_completion_length": 53, "max_completion_length": 272, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06029437027697187, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 9.856643438339233, "avg_step_time_sec": 13.125903831111442, "gpu_allocated_gb": 32.438163968, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1826.3498063087463, "eta_sec": 800.6801336977979, "timestamp": "2026-02-15T17:27:01.780215"}
{"step": 140, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.74, "mean_correctness_reward": 0.74, "mean_length_penalty": -0.105986328125, "mean_total_reward": 0.73682041015625, "std_total_reward": 0.43895162268666854, "total_tokens": 10853, "mean_completion_length": 217.06, "median_completion_length": 185, "min_completion_length": 74, "max_completion_length": 1019, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.14042690876071903, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 3.7163245677948, "avg_step_time_sec": 13.05869255065918, "gpu_allocated_gb": 32.437270016, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1830.073839187622, "eta_sec": 783.5215530395508, "timestamp": "2026-02-15T17:27:10.022672"}
{"step": 141, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.10189453125, "mean_total_reward": 0.6569431640625, "std_total_reward": 0.47452436869943904, "total_tokens": 10434, "mean_completion_length": 208.68, "median_completion_length": 117, "min_completion_length": 36, "max_completion_length": 1362, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15824264205131602, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 11.028348922729492, "avg_step_time_sec": 13.044292950461097, "gpu_allocated_gb": 32.438350336, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1841.1095016002655, "eta_sec": 769.6132840772048, "timestamp": "2026-02-15T17:27:23.021036"}
{"step": 142, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.050791015625, "mean_total_reward": 0.79847626953125, "std_total_reward": 0.39998043066245126, "total_tokens": 5201, "mean_completion_length": 104.02, "median_completion_length": 93, "min_completion_length": 56, "max_completion_length": 186, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.00022092138149347107, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 14.231054544448853, "avg_step_time_sec": 13.052650426475095, "gpu_allocated_gb": 32.438835712, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1855.3484892845154, "eta_sec": 757.0537247355555, "timestamp": "2026-02-15T17:27:30.052540"}
{"step": 143, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.62, "mean_correctness_reward": 0.62, "mean_length_penalty": -0.072294921875, "mean_total_reward": 0.61783115234375, "std_total_reward": 0.48523018697050296, "total_tokens": 7403, "mean_completion_length": 148.06, "median_completion_length": 135, "min_completion_length": 70, "max_completion_length": 281, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2517474489090201, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.773440361022949, "avg_step_time_sec": 12.9807678386048, "gpu_allocated_gb": 32.437147136, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1858.1293277740479, "eta_sec": 739.9037668004736, "timestamp": "2026-02-15T17:27:33.547295"}
{"step": 144, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.1849609375, "mean_total_reward": 0.694451171875, "std_total_reward": 0.45831095575717284, "total_tokens": 18940, "mean_completion_length": 378.8, "median_completion_length": 351, "min_completion_length": 146, "max_completion_length": 1072, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15881731852260153, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 3.7976038455963135, "avg_step_time_sec": 12.91699586643113, "gpu_allocated_gb": 32.437300224, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1861.9355850219727, "eta_sec": 723.3517685201432, "timestamp": "2026-02-15T17:27:42.769246"}
{"step": 145, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.070732421875, "mean_total_reward": 0.79787802734375, "std_total_reward": 0.3997874628958622, "total_tokens": 7243, "mean_completion_length": 144.86, "median_completion_length": 125, "min_completion_length": 75, "max_completion_length": 357, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.00039622127926071, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 12.395339488983154, "avg_step_time_sec": 12.913398236241834, "gpu_allocated_gb": 32.438395392, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1874.3381490707397, "eta_sec": 710.2369029933009, "timestamp": "2026-02-15T17:27:50.149240"}
{"step": 146, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.9, "mean_correctness_reward": 0.9, "mean_length_penalty": -0.05134765625, "mean_total_reward": 0.8984595703125, "std_total_reward": 0.30014554550605815, "total_tokens": 5258, "mean_completion_length": 105.16, "median_completion_length": 90, "min_completion_length": 46, "max_completion_length": 253, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1581424260562514, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.3548853397369385, "avg_step_time_sec": 12.85477828489591, "gpu_allocated_gb": 32.437380096, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1878.700411081314, "eta_sec": 694.1580273843791, "timestamp": "2026-02-15T17:27:53.758879"}
{"step": 147, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.36, "mean_correctness_reward": 0.36, "mean_length_penalty": -0.061083984375, "mean_total_reward": 0.35816748046875, "std_total_reward": 0.4805171260562829, "total_tokens": 6255, "mean_completion_length": 125.1, "median_completion_length": 108, "min_completion_length": 41, "max_completion_length": 384, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0804139037812783, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 3.230955123901367, "avg_step_time_sec": 12.78931010012724, "gpu_allocated_gb": 32.437225984, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1881.9389066696167, "eta_sec": 677.8334353067437, "timestamp": "2026-02-15T17:27:57.830281"}
{"step": 148, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.52, "mean_correctness_reward": 0.52, "mean_length_penalty": -0.0817578125, "mean_total_reward": 0.517547265625, "std_total_reward": 0.4993644223334874, "total_tokens": 8372, "mean_completion_length": 167.44, "median_completion_length": 137, "min_completion_length": 105, "max_completion_length": 441, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.23164521672570393, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 4.551592826843262, "avg_step_time_sec": 12.733649848280725, "gpu_allocated_gb": 32.437413888, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1886.497836112976, "eta_sec": 662.1497921105977, "timestamp": "2026-02-15T17:28:02.865749"}
{"step": 149, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.106884765625, "mean_total_reward": 0.57679345703125, "std_total_reward": 0.49379181602838323, "total_tokens": 10945, "mean_completion_length": 218.9, "median_completion_length": 139, "min_completion_length": 58, "max_completion_length": 1053, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0609897449340149, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 5.250110149383545, "avg_step_time_sec": 12.68342474963041, "gpu_allocated_gb": 32.437510656, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1891.755053281784, "eta_sec": 646.8546622311509, "timestamp": "2026-02-15T17:28:11.847230"}
{"step": 150, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.54, "mean_correctness_reward": 0.54, "mean_length_penalty": -0.22123046875, "mean_total_reward": 0.5333630859375, "std_total_reward": 0.504398838542082, "total_tokens": 22654, "mean_completion_length": 453.08, "median_completion_length": 130, "min_completion_length": 21, "max_completion_length": 2048, "frac_has_think": 0.88, "frac_has_boxed": 0.88, "mean_reward_std_per_prompt": 0.09290628442998113, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 11.39000129699707, "avg_step_time_sec": 12.674801926612854, "gpu_allocated_gb": 32.4384256, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1903.1524748802185, "eta_sec": 633.7400963306427, "timestamp": "2026-02-15T17:28:30.736015"}
{"step": 151, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.0688671875, "mean_total_reward": 0.957933984375, "std_total_reward": 0.19589292093136892, "total_tokens": 7052, "mean_completion_length": 141.04, "median_completion_length": 145, "min_completion_length": 53, "max_completion_length": 310, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12031316626600434, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.349634170532227, "avg_step_time_sec": 12.738873663327551, "gpu_allocated_gb": 32.439773184, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1925.858648777008, "eta_sec": 624.20480950305, "timestamp": "2026-02-15T17:28:41.349443"}
{"step": 152, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.9, "mean_correctness_reward": 0.9, "mean_length_penalty": -0.0754296875, "mean_total_reward": 0.897737109375, "std_total_reward": 0.3002405567767182, "total_tokens": 7724, "mean_completion_length": 154.48, "median_completion_length": 159, "min_completion_length": 88, "max_completion_length": 349, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2601389755899099, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.2304298877716064, "avg_step_time_sec": 12.682897059541, "gpu_allocated_gb": 32.437312512, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1930.0966856479645, "eta_sec": 608.779058857968, "timestamp": "2026-02-15T17:28:45.861698"}
{"step": 153, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.0703515625, "mean_total_reward": 0.597889453125, "std_total_reward": 0.4901285757549346, "total_tokens": 7204, "mean_completion_length": 144.08, "median_completion_length": 127, "min_completion_length": 54, "max_completion_length": 277, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0003776622246786812, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 4.321904420852661, "avg_step_time_sec": 12.628250048830619, "gpu_allocated_gb": 32.43737856, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1934.426521539688, "eta_sec": 593.5277522950391, "timestamp": "2026-02-15T17:28:49.810441"}
{"step": 154, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.3, "mean_correctness_reward": 0.3, "mean_length_penalty": -0.12703125, "mean_total_reward": 0.2961890625, "std_total_reward": 0.4583789253745462, "total_tokens": 13008, "mean_completion_length": 260.16, "median_completion_length": 159, "min_completion_length": 77, "max_completion_length": 1583, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.10131524135338256, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 3.7752845287323, "avg_step_time_sec": 12.57076325973907, "gpu_allocated_gb": 32.437288448, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1938.2093949317932, "eta_sec": 578.2551099479972, "timestamp": "2026-02-15T17:29:01.692648"}
{"step": 155, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.05951171875, "mean_total_reward": 0.7982146484375, "std_total_reward": 0.39986450399491513, "total_tokens": 6094, "mean_completion_length": 121.88, "median_completion_length": 106, "min_completion_length": 33, "max_completion_length": 658, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.0005152158723967431, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 17.300174951553345, "avg_step_time_sec": 12.601275593234647, "gpu_allocated_gb": 32.43913984, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1955.518545627594, "eta_sec": 567.0574016955591, "timestamp": "2026-02-15T17:29:13.117855"}
{"step": 156, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.5, "mean_correctness_reward": 0.5, "mean_length_penalty": -0.037548828125, "mean_total_reward": 0.49887353515625, "std_total_reward": 0.4999047123608195, "total_tokens": 3845, "mean_completion_length": 76.9, "median_completion_length": 64, "min_completion_length": 18, "max_completion_length": 221, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.26955167145092557, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 7.8832807540893555, "avg_step_time_sec": 12.57103203657346, "gpu_allocated_gb": 32.438001664, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1963.4105603694916, "eta_sec": 553.1254096092322, "timestamp": "2026-02-15T17:29:18.199067"}
{"step": 157, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.18982421875, "mean_total_reward": 0.7743052734375, "std_total_reward": 0.4198181751396815, "total_tokens": 19438, "mean_completion_length": 388.76, "median_completion_length": 174, "min_completion_length": 99, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.1802838115327148, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.934600591659546, "avg_step_time_sec": 12.50965349233834, "gpu_allocated_gb": 32.437179904, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1966.3527636528015, "eta_sec": 537.9151001705486, "timestamp": "2026-02-15T17:29:33.261965"}
{"step": 158, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.52, "mean_correctness_reward": 0.52, "mean_length_penalty": -0.216123046875, "mean_total_reward": 0.51351630859375, "std_total_reward": 0.49995563064169063, "total_tokens": 22131, "mean_completion_length": 442.62, "median_completion_length": 205, "min_completion_length": 50, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.96, "mean_reward_std_per_prompt": 0.33043343771263195, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.007426738739014, "avg_step_time_sec": 12.569765981239609, "gpu_allocated_gb": 32.439789568, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 1988.3687679767609, "eta_sec": 527.9301712120636, "timestamp": "2026-02-15T17:29:55.481140"}
{"step": 159, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.06208984375, "mean_total_reward": 0.7981373046875, "std_total_reward": 0.40126961043169507, "total_tokens": 6358, "mean_completion_length": 127.16, "median_completion_length": 93, "min_completion_length": 27, "max_completion_length": 622, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1201706619538279, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.69644832611084, "avg_step_time_sec": 12.633455807307982, "gpu_allocated_gb": 32.439805952, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2011.0728089809418, "eta_sec": 517.9716880996273, "timestamp": "2026-02-15T17:30:08.325083"}
{"step": 160, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.86, "mean_correctness_reward": 0.86, "mean_length_penalty": -0.160986328125, "mean_total_reward": 0.85517041015625, "std_total_reward": 0.3509429218315006, "total_tokens": 16485, "mean_completion_length": 329.7, "median_completion_length": 137, "min_completion_length": 59, "max_completion_length": 2048, "frac_has_think": 0.98, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.15845749281107158, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 6.698280096054077, "avg_step_time_sec": 12.596360959112644, "gpu_allocated_gb": 32.437750272, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2017.7791123390198, "eta_sec": 503.85443836450577, "timestamp": "2026-02-15T17:30:24.135023"}
{"step": 161, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.36, "mean_correctness_reward": 0.36, "mean_length_penalty": -0.048037109375, "mean_total_reward": 0.35855888671875, "std_total_reward": 0.47974607373113426, "total_tokens": 4919, "mean_completion_length": 98.38, "median_completion_length": 93, "min_completion_length": 37, "max_completion_length": 146, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.08012166280094972, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 21.371121406555176, "avg_step_time_sec": 12.650862576798623, "gpu_allocated_gb": 32.439758848, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2039.1582720279694, "eta_sec": 493.3836404951463, "timestamp": "2026-02-15T17:30:33.631423"}
{"step": 162, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.38, "mean_correctness_reward": 0.38, "mean_length_penalty": -0.263408203125, "mean_total_reward": 0.37209775390625, "std_total_reward": 0.48980299808630007, "total_tokens": 26973, "mean_completion_length": 539.46, "median_completion_length": 152, "min_completion_length": 61, "max_completion_length": 2048, "frac_has_think": 0.84, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.2604455069179408, "frac_prompts_all_correct": 0.0, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 2.5931737422943115, "avg_step_time_sec": 12.588778077820201, "gpu_allocated_gb": 32.437110784, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2041.7588176727295, "eta_sec": 478.37356695716767, "timestamp": "2026-02-15T17:30:49.428844"}
{"step": 163, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.169033203125, "mean_total_reward": 0.65492900390625, "std_total_reward": 0.4768314609007365, "total_tokens": 17309, "mean_completion_length": 346.18, "median_completion_length": 124, "min_completion_length": 53, "max_completion_length": 2048, "frac_has_think": 0.92, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.15933149655266407, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.741898775100708, "avg_step_time_sec": 12.651067161852596, "gpu_allocated_gb": 32.439774208, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2064.507397890091, "eta_sec": 468.0894849885461, "timestamp": "2026-02-15T17:31:11.321936"}
{"step": 164, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.05607421875, "mean_total_reward": 0.6583177734375, "std_total_reward": 0.47359221498874554, "total_tokens": 5742, "mean_completion_length": 114.84, "median_completion_length": 90, "min_completion_length": 44, "max_completion_length": 325, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1896873195781884, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 21.691614627838135, "avg_step_time_sec": 12.706192451279337, "gpu_allocated_gb": 32.43973632, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2086.207038164139, "eta_sec": 457.42292824605613, "timestamp": "2026-02-15T17:31:21.811029"}
{"step": 165, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.052861328125, "mean_total_reward": 0.95841416015625, "std_total_reward": 0.19599620886048272, "total_tokens": 5413, "mean_completion_length": 108.26, "median_completion_length": 82, "min_completion_length": 26, "max_completion_length": 374, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.12003175680997578, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.675021171569824, "avg_step_time_sec": 12.657518685947766, "gpu_allocated_gb": 32.437335552, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2090.890193462372, "eta_sec": 443.0131540081718, "timestamp": "2026-02-15T17:31:26.712754"}
{"step": 166, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.88, "mean_correctness_reward": 0.88, "mean_length_penalty": -0.073828125, "mean_total_reward": 0.87778515625, "std_total_reward": 0.3253298209160685, "total_tokens": 7560, "mean_completion_length": 151.2, "median_completion_length": 129, "min_completion_length": 51, "max_completion_length": 410, "frac_has_think": 1.0, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.1782627627836837, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.420958995819092, "avg_step_time_sec": 12.607900856489158, "gpu_allocated_gb": 32.43739392, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2095.3186008930206, "eta_sec": 428.66862912063135, "timestamp": "2026-02-15T17:31:31.597854"}
{"step": 167, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.0998828125, "mean_total_reward": 0.757003515625, "std_total_reward": 0.4306094089453257, "total_tokens": 10228, "mean_completion_length": 204.56, "median_completion_length": 61, "min_completion_length": 33, "max_completion_length": 1936, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.15196446871918445, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 5.109460830688477, "avg_step_time_sec": 12.56300001801131, "gpu_allocated_gb": 32.437479936, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2100.438188314438, "eta_sec": 414.57900059437327, "timestamp": "2026-02-15T17:31:45.829639"}
{"step": 168, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.94, "mean_correctness_reward": 0.94, "mean_length_penalty": -0.0511328125, "mean_total_reward": 0.938466015625, "std_total_reward": 0.23778471677673646, "total_tokens": 5236, "mean_completion_length": 104.72, "median_completion_length": 99, "min_completion_length": 36, "max_completion_length": 253, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.14028501257548046, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 19.908775568008423, "avg_step_time_sec": 12.606724872475578, "gpu_allocated_gb": 32.439611904, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2120.3572051525116, "eta_sec": 403.4151959192185, "timestamp": "2026-02-15T17:31:55.400363"}
{"step": 169, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.185751953125, "mean_total_reward": 0.47442744140625, "std_total_reward": 0.5032306130004146, "total_tokens": 19021, "mean_completion_length": 380.42, "median_completion_length": 158, "min_completion_length": 82, "max_completion_length": 1997, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.16257099264983618, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 3.2178640365600586, "avg_step_time_sec": 12.551169482913949, "gpu_allocated_gb": 32.437220864, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2123.583566904068, "eta_sec": 389.0862539703324, "timestamp": "2026-02-15T17:32:10.092215"}
{"step": 170, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.88, "mean_correctness_reward": 0.88, "mean_length_penalty": -0.05890625, "mean_total_reward": 0.8782328125, "std_total_reward": 0.32594595114681246, "total_tokens": 6032, "mean_completion_length": 120.64, "median_completion_length": 100, "min_completion_length": 44, "max_completion_length": 692, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.1837885485210364, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 21.4967622756958, "avg_step_time_sec": 12.603790616989135, "gpu_allocated_gb": 32.439735808, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2145.088668823242, "eta_sec": 378.1137185096741, "timestamp": "2026-02-15T17:32:22.678695"}
{"step": 171, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.08515625, "mean_total_reward": 0.7774453125, "std_total_reward": 0.41594138648025153, "total_tokens": 8720, "mean_completion_length": 174.4, "median_completion_length": 120, "min_completion_length": 36, "max_completion_length": 675, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.06057757064828011, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 7.235921859741211, "avg_step_time_sec": 12.572399571625112, "gpu_allocated_gb": 32.437846528, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2152.3331248760223, "eta_sec": 364.59958757712826, "timestamp": "2026-02-15T17:32:30.126874"}
{"step": 172, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.66, "mean_correctness_reward": 0.66, "mean_length_penalty": -0.120771484375, "mean_total_reward": 0.65637685546875, "std_total_reward": 0.47442868281896583, "total_tokens": 12367, "mean_completion_length": 247.34, "median_completion_length": 98, "min_completion_length": 44, "max_completion_length": 1889, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2999328563738762, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.4296886920928955, "avg_step_time_sec": 12.542500089767367, "gpu_allocated_gb": 32.437833728, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2159.7715542316437, "eta_sec": 351.19000251348626, "timestamp": "2026-02-15T17:32:44.892906"}
{"step": 173, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.057626953125, "mean_total_reward": 0.79827119140625, "std_total_reward": 0.40027058394815024, "total_tokens": 5901, "mean_completion_length": 118.02, "median_completion_length": 104, "min_completion_length": 51, "max_completion_length": 291, "frac_has_think": 1.0, "frac_has_boxed": 0.98, "mean_reward_std_per_prompt": 0.24970209446887254, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 19.80413031578064, "avg_step_time_sec": 12.584474830958195, "gpu_allocated_gb": 32.439555072, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2179.585020303726, "eta_sec": 339.78082043587125, "timestamp": "2026-02-15T17:32:54.950026"}
{"step": 174, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.072314453125, "mean_total_reward": 0.47783056640625, "std_total_reward": 0.5001041357252412, "total_tokens": 7405, "mean_completion_length": 148.1, "median_completion_length": 92, "min_completion_length": 44, "max_completion_length": 2048, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.09865750654294385, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 3.821180582046509, "avg_step_time_sec": 12.53411107090698, "gpu_allocated_gb": 32.437316608, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2183.414259672165, "eta_sec": 325.88688784358146, "timestamp": "2026-02-15T17:33:08.567013"}
{"step": 175, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.68, "mean_correctness_reward": 0.68, "mean_length_penalty": -0.155537109375, "mean_total_reward": 0.67533388671875, "std_total_reward": 0.46830732237476413, "total_tokens": 15927, "mean_completion_length": 318.54, "median_completion_length": 211, "min_completion_length": 56, "max_completion_length": 1791, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.2183582334296339, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 20.46109938621521, "avg_step_time_sec": 12.579408146994455, "gpu_allocated_gb": 32.439810048, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2203.8850934505463, "eta_sec": 314.4852036748614, "timestamp": "2026-02-15T17:33:28.498160"}
{"step": 176, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.05484375, "mean_total_reward": 0.9983546875, "std_total_reward": 0.0006942521211277848, "total_tokens": 5616, "mean_completion_length": 112.32, "median_completion_length": 94, "min_completion_length": 51, "max_completion_length": 249, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.00024234510247840812, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 18.54299283027649, "avg_step_time_sec": 12.613292150876738, "gpu_allocated_gb": 32.439410688, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2222.4372029304504, "eta_sec": 302.7190116210417, "timestamp": "2026-02-15T17:33:37.522818"}
{"step": 177, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.6, "mean_correctness_reward": 0.6, "mean_length_penalty": -0.1484375, "mean_total_reward": 0.595546875, "std_total_reward": 0.4923968636102983, "total_tokens": 15200, "mean_completion_length": 304, "median_completion_length": 152, "min_completion_length": 75, "max_completion_length": 1511, "frac_has_think": 1.0, "frac_has_boxed": 1.0, "mean_reward_std_per_prompt": 0.23278790888316558, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 3.308896780014038, "avg_step_time_sec": 12.56072494539164, "gpu_allocated_gb": 32.437240832, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2225.753970384598, "eta_sec": 288.8966737440077, "timestamp": "2026-02-15T17:33:48.923308"}
{"step": 178, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.42, "mean_correctness_reward": 0.42, "mean_length_penalty": -0.248564453125, "mean_total_reward": 0.41254306640625, "std_total_reward": 0.49876757763521584, "total_tokens": 25453, "mean_completion_length": 509.06, "median_completion_length": 152, "min_completion_length": 57, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.0627976016048749, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 16.050081968307495, "avg_step_time_sec": 12.580328074733863, "gpu_allocated_gb": 32.439040512, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2241.8114733695984, "eta_sec": 276.767217644145, "timestamp": "2026-02-15T17:34:09.321920"}
{"step": 179, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.207861328125, "mean_total_reward": 0.39376416015625, "std_total_reward": 0.4943222379288378, "total_tokens": 21285, "mean_completion_length": 425.7, "median_completion_length": 145, "min_completion_length": 39, "max_completion_length": 2048, "frac_has_think": 0.86, "frac_has_boxed": 0.86, "mean_reward_std_per_prompt": 0.002133657748044701, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.6, "step_time_sec": 22.86830496788025, "avg_step_time_sec": 12.63780280598049, "gpu_allocated_gb": 32.439779328, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2264.687174320221, "eta_sec": 265.3938589255903, "timestamp": "2026-02-15T17:34:31.801227"}
{"step": 180, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.05798828125, "mean_total_reward": 0.7582603515625, "std_total_reward": 0.4271589267532032, "total_tokens": 5938, "mean_completion_length": 118.76, "median_completion_length": 117, "min_completion_length": 54, "max_completion_length": 277, "frac_has_think": 1.0, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.20020636227614183, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 22.177975177764893, "avg_step_time_sec": 12.690803763601515, "gpu_allocated_gb": 32.439780352, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2286.8730022907257, "eta_sec": 253.8160752720303, "timestamp": "2026-02-15T17:34:42.111947"}
{"step": 181, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.84, "mean_correctness_reward": 0.84, "mean_length_penalty": -0.074521484375, "mean_total_reward": 0.8377643554687499, "std_total_reward": 0.36732048920425764, "total_tokens": 7631, "mean_completion_length": 152.62, "median_completion_length": 142, "min_completion_length": 50, "max_completion_length": 435, "frac_has_think": 1.0, "frac_has_boxed": 0.94, "mean_reward_std_per_prompt": 0.21815484406704455, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 3.5108344554901123, "avg_step_time_sec": 12.640085701125761, "gpu_allocated_gb": 32.437262848, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2290.3914771080017, "eta_sec": 240.16162832138946, "timestamp": "2026-02-15T17:34:46.755988"}
{"step": 182, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.98, "mean_correctness_reward": 0.98, "mean_length_penalty": -0.03818359375, "mean_total_reward": 0.9788544921875, "std_total_reward": 0.13992716111038755, "total_tokens": 3910, "mean_completion_length": 78.2, "median_completion_length": 72, "min_completion_length": 37, "max_completion_length": 192, "frac_has_think": 1.0, "frac_has_boxed": 0.92, "mean_reward_std_per_prompt": 0.06019408379781807, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 5.354402303695679, "avg_step_time_sec": 12.600054473667354, "gpu_allocated_gb": 32.437540864, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2295.7548112869263, "eta_sec": 226.80098052601238, "timestamp": "2026-02-15T17:34:50.370325"}
{"step": 183, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.05169921875, "mean_total_reward": 0.9984490234375, "std_total_reward": 0.0005386693964092682, "total_tokens": 5294, "mean_completion_length": 105.88, "median_completion_length": 106, "min_completion_length": 27, "max_completion_length": 192, "frac_has_think": 1.0, "frac_has_boxed": 0.74, "mean_reward_std_per_prompt": 0.0002798823889439144, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.6926980018615723, "avg_step_time_sec": 12.54591591371213, "gpu_allocated_gb": 32.437157376, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2298.456084251404, "eta_sec": 213.28057053310621, "timestamp": "2026-02-15T17:34:53.165960"}
{"step": 184, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.4, "mean_correctness_reward": 0.4, "mean_length_penalty": -0.0679296875, "mean_total_reward": 0.397962109375, "std_total_reward": 0.4908323307825718, "total_tokens": 6956, "mean_completion_length": 139.12, "median_completion_length": 132, "min_completion_length": 19, "max_completion_length": 1139, "frac_has_think": 1.0, "frac_has_boxed": 0.84, "mean_reward_std_per_prompt": 0.12103150116744718, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.4, "step_time_sec": 2.757025718688965, "avg_step_time_sec": 12.492715423521789, "gpu_allocated_gb": 32.437140992, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2301.220032453537, "eta_sec": 199.88344677634862, "timestamp": "2026-02-15T17:35:01.333217"}
{"step": 185, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.86, "mean_correctness_reward": 0.86, "mean_length_penalty": -0.067119140625, "mean_total_reward": 0.85798642578125, "std_total_reward": 0.34756259195043576, "total_tokens": 6873, "mean_completion_length": 137.46, "median_completion_length": 129, "min_completion_length": 25, "max_completion_length": 533, "frac_has_think": 1.0, "frac_has_boxed": 0.6, "mean_reward_std_per_prompt": 0.15815986991008138, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 11.817285537719727, "avg_step_time_sec": 12.4890644511661, "gpu_allocated_gb": 32.438574592, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2313.044615983963, "eta_sec": 187.33596676749153, "timestamp": "2026-02-15T17:35:09.788892"}
{"step": 186, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.037177734375, "mean_total_reward": 0.99888466796875, "std_total_reward": 0.0004909041758493478, "total_tokens": 3807, "mean_completion_length": 76.14, "median_completion_length": 80, "min_completion_length": 24, "max_completion_length": 151, "frac_has_think": 1.0, "frac_has_boxed": 0.7, "mean_reward_std_per_prompt": 0.0002295244956925337, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 5.911600589752197, "avg_step_time_sec": 12.453701742233768, "gpu_allocated_gb": 32.437635584, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2318.965842962265, "eta_sec": 174.35182439127277, "timestamp": "2026-02-15T17:35:13.515645"}
{"step": 187, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.203134765625, "mean_total_reward": 0.79390595703125, "std_total_reward": 0.40834572057018337, "total_tokens": 20801, "mean_completion_length": 416.02, "median_completion_length": 165, "min_completion_length": 58, "max_completion_length": 2048, "frac_has_think": 0.92, "frac_has_boxed": 0.46, "mean_reward_std_per_prompt": 0.0019083201107899072, "frac_prompts_all_correct": 0.8, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.8393442630767822, "avg_step_time_sec": 12.40228806587464, "gpu_allocated_gb": 32.437216768, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2321.8127479553223, "eta_sec": 161.2297448563703, "timestamp": "2026-02-15T17:35:29.277202"}
{"step": 188, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.76, "mean_correctness_reward": 0.76, "mean_length_penalty": -0.08712890625, "mean_total_reward": 0.7573861328125, "std_total_reward": 0.4268456786203832, "total_tokens": 8922, "mean_completion_length": 178.44, "median_completion_length": 113, "min_completion_length": 78, "max_completion_length": 703, "frac_has_think": 1.0, "frac_has_boxed": 0.72, "mean_reward_std_per_prompt": 0.12012881016857876, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 22.54631781578064, "avg_step_time_sec": 12.456245670927332, "gpu_allocated_gb": 32.439775232, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2344.3668310642242, "eta_sec": 149.474948051128, "timestamp": "2026-02-15T17:35:42.340172"}
{"step": 189, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.049765625, "mean_total_reward": 0.99850703125, "std_total_reward": 0.0007976008441748498, "total_tokens": 5096, "mean_completion_length": 101.92, "median_completion_length": 85, "min_completion_length": 44, "max_completion_length": 239, "frac_has_think": 1.0, "frac_has_boxed": 0.18, "mean_reward_std_per_prompt": 0.00023827732903343964, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 7.836302042007446, "avg_step_time_sec": 12.431801524742571, "gpu_allocated_gb": 32.437888512, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2352.21119761467, "eta_sec": 136.7498167721683, "timestamp": "2026-02-15T17:35:47.235600"}
{"step": 190, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.96, "mean_correctness_reward": 0.96, "mean_length_penalty": -0.071298828125, "mean_total_reward": 0.95786103515625, "std_total_reward": 0.19615569321606086, "total_tokens": 7301, "mean_completion_length": 146.02, "median_completion_length": 93, "min_completion_length": 44, "max_completion_length": 499, "frac_has_think": 1.0, "frac_has_boxed": 0.26, "mean_reward_std_per_prompt": 0.12017687624009166, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 3.246194362640381, "avg_step_time_sec": 12.383456223889402, "gpu_allocated_gb": 32.437226496, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2355.465350151062, "eta_sec": 123.83456223889402, "timestamp": "2026-02-15T17:35:52.094193"}
{"step": 191, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.7, "mean_correctness_reward": 0.7, "mean_length_penalty": -0.053330078125, "mean_total_reward": 0.69840009765625, "std_total_reward": 0.45871622365421527, "total_tokens": 5461, "mean_completion_length": 109.22, "median_completion_length": 105, "min_completion_length": 37, "max_completion_length": 285, "frac_has_think": 1.0, "frac_has_boxed": 0.1, "mean_reward_std_per_prompt": 0.10033684081576517, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 5.643501043319702, "avg_step_time_sec": 12.348168500430921, "gpu_allocated_gb": 32.437571072, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2361.117472410202, "eta_sec": 111.13351650387828, "timestamp": "2026-02-15T17:35:56.392135"}
{"step": 192, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.78, "mean_correctness_reward": 0.78, "mean_length_penalty": -0.046455078125, "mean_total_reward": 0.77860634765625, "std_total_reward": 0.4139994308634661, "total_tokens": 4757, "mean_completion_length": 95.14, "median_completion_length": 95, "min_completion_length": 42, "max_completion_length": 169, "frac_has_think": 1.0, "frac_has_boxed": 0.08, "mean_reward_std_per_prompt": 0.060109680453779374, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 3.6132733821868896, "avg_step_time_sec": 12.3026742550234, "gpu_allocated_gb": 32.437272576, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2364.7374217510223, "eta_sec": 98.4213940401872, "timestamp": "2026-02-15T17:35:59.298779"}
{"step": 193, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.48, "mean_correctness_reward": 0.48, "mean_length_penalty": -0.06259765625, "mean_total_reward": 0.4781220703125, "std_total_reward": 0.5000841186650551, "total_tokens": 6410, "mean_completion_length": 128.2, "median_completion_length": 109, "min_completion_length": 47, "max_completion_length": 447, "frac_has_think": 1.0, "frac_has_boxed": 0.08, "mean_reward_std_per_prompt": 0.2580999617698539, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 2.5950205326080322, "avg_step_time_sec": 12.252375531073062, "gpu_allocated_gb": 32.437111296, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2367.3426127433777, "eta_sec": 85.76662871751144, "timestamp": "2026-02-15T17:36:03.651832"}
{"step": 194, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.64, "mean_correctness_reward": 0.64, "mean_length_penalty": -0.15830078125, "mean_total_reward": 0.6352509765625, "std_total_reward": 0.48389511930815093, "total_tokens": 16210, "mean_completion_length": 324.2, "median_completion_length": 131, "min_completion_length": 39, "max_completion_length": 1534, "frac_has_think": 1.0, "frac_has_boxed": 0.16, "mean_reward_std_per_prompt": 0.33020565776533634, "frac_prompts_all_correct": 0.2, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 5.206594944000244, "avg_step_time_sec": 12.216057074438666, "gpu_allocated_gb": 32.437512704, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2372.5594124794006, "eta_sec": 73.296342446632, "timestamp": "2026-02-15T17:36:16.075271"}
{"step": 195, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.8, "mean_correctness_reward": 0.8, "mean_length_penalty": -0.0426171875, "mean_total_reward": 0.798721484375, "std_total_reward": 0.40040156306081737, "total_tokens": 4364, "mean_completion_length": 87.28, "median_completion_length": 42, "min_completion_length": 12, "max_completion_length": 331, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.19984647251195548, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 16.815892934799194, "avg_step_time_sec": 12.239645976286669, "gpu_allocated_gb": 32.439036928, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2389.3837225437164, "eta_sec": 61.19822988143334, "timestamp": "2026-02-15T17:36:24.906068"}
{"step": 196, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.9, "mean_correctness_reward": 0.9, "mean_length_penalty": -0.039765625, "mean_total_reward": 0.8988070312499999, "std_total_reward": 0.29988617011354113, "total_tokens": 4072, "mean_completion_length": 81.44, "median_completion_length": 61, "min_completion_length": 28, "max_completion_length": 222, "frac_has_think": 1.0, "frac_has_boxed": 0.02, "mean_reward_std_per_prompt": 0.15814293313720051, "frac_prompts_all_correct": 0.6, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 4.071849822998047, "avg_step_time_sec": 12.197973546933154, "gpu_allocated_gb": 32.437351424, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2393.464239358902, "eta_sec": 48.791894187732616, "timestamp": "2026-02-15T17:36:28.274625"}
{"step": 197, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.040341796875, "mean_total_reward": 0.99878974609375, "std_total_reward": 0.00044796339752330833, "total_tokens": 4131, "mean_completion_length": 82.62, "median_completion_length": 77, "min_completion_length": 27, "max_completion_length": 145, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.00020512232446097826, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.995314598083496, "avg_step_time_sec": 12.151259542116659, "gpu_allocated_gb": 32.437213696, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2396.4674532413483, "eta_sec": 36.45377862634997, "timestamp": "2026-02-15T17:36:30.846189"}
{"step": 198, "num_prompts": 5, "num_rollouts": 50, "accuracy": 1.0, "mean_correctness_reward": 1.0, "mean_length_penalty": -0.02095703125, "mean_total_reward": 0.9993712890625001, "std_total_reward": 0.00021209076014657113, "total_tokens": 2146, "mean_completion_length": 42.92, "median_completion_length": 43, "min_completion_length": 16, "max_completion_length": 69, "frac_has_think": 1.0, "frac_has_boxed": 0.0, "mean_reward_std_per_prompt": 0.00014366592553992385, "frac_prompts_all_correct": 1.0, "frac_prompts_all_incorrect": 0.0, "step_time_sec": 2.3903262615203857, "avg_step_time_sec": 12.101961899285365, "gpu_allocated_gb": 32.437084672, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2398.865796804428, "eta_sec": 24.20392379857073, "timestamp": "2026-02-15T17:36:32.680520"}
{"step": 199, "num_prompts": 5, "num_rollouts": 50, "accuracy": 0.58, "mean_correctness_reward": 0.58, "mean_length_penalty": -0.08693359375, "mean_total_reward": 0.5773919921875, "std_total_reward": 0.49473569587997923, "total_tokens": 8902, "mean_completion_length": 178.04, "median_completion_length": 171, "min_completion_length": 34, "max_completion_length": 823, "frac_has_think": 1.0, "frac_has_boxed": 0.28, "mean_reward_std_per_prompt": 0.19857907356781085, "frac_prompts_all_correct": 0.4, "frac_prompts_all_incorrect": 0.2, "step_time_sec": 1.8475439548492432, "avg_step_time_sec": 12.050432160871113, "gpu_allocated_gb": 32.436967936, "gpu_reserved_gb": 81.197531136, "gpu_peak_gb": 74.331317248, "elapsed_sec": 2400.7206428050995, "eta_sec": 12.050432160871113, "timestamp": "2026-02-15T17:36:39.545156"}