healed/grid_math/glean_keep25_s1226.console.log

/home/henry/Documents/PythonProjects/variable-reap/.venv/lib/python3.12/site-packages/megablocks/grouped_gemm_util.py:10: UserWarning: Grouped GEMM not available.
  warnings.warn('Grouped GEMM not available.')
wandb: [wandb.login()] Loaded credentials for https://api.wandb.ai from /home/henry/.netrc.
wandb: Currently logged in as: hbfreed to https://api.wandb.ai. Use `wandb login --relogin` to force relogin
wandb: setting up run vhgtf0ej
wandb: Tracking run with wandb version 0.28.0
wandb: Run data is saved locally in outputs/healed/grid_math/glean_keep25_s1226/wandb/run-20260716_034735-vhgtf0ej
wandb: Run `wandb offline` to turn off syncing.
wandb: Syncing run glean-math-keep25-s1226
wandb: ⭐️ View project at https://wandb.ai/hbfreed/glean-grid
wandb: 🚀 View run at https://wandb.ai/hbfreed/glean-grid/runs/vhgtf0ej
12115 cached top-128 chat trajectories / 6,476,634 unique tokens | 53 steps/epoch | 150 total steps | student params 2.09B | teacher overlap=False
{"step": 1, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 1.4740517740617196, "tokens": 120000, "cumulative_loss_tokens": 120000, "grad_norm": 121.0, "lr": 6e-06, "finish_rate": 0.902, "comp_len": 472.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 41.0, "frames": {"chat": 254}, "mem_gb": 9.82}
The attention mask is not set and cannot be inferred from input because pad token is same as eos token. As a consequence, you may observe unexpected behavior. Please pass your input's `attention_mask` to obtain reliable results.
[eval step 1] sample: 'The value is 28, and the perimeter of the resulting triangle is 28. The problem is to find the perimeter of the resulting triangle, given the conditions mentioned.\n\nThe perimeter of the resulting tria'
{"step": 2, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 1.377854461752375, "tokens": 120000, "cumulative_loss_tokens": 240000, "grad_norm": 95.0, "lr": 9e-06, "finish_rate": 0.876, "comp_len": 497.9, "t_data_s": 0.1, "t_rollout_s": 0.0, "t_step_s": 33.9, "frames": {"chat": 241}, "mem_gb": 9.98}
{"step": 3, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 1.2003990252320964, "tokens": 120000, "cumulative_loss_tokens": 360000, "grad_norm": 59.75, "lr": 1.2e-05, "finish_rate": 0.746, "comp_len": 563.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.3, "frames": {"chat": 213}, "mem_gb": 10.01}
{"step": 4, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.8684858515068888, "tokens": 120000, "cumulative_loss_tokens": 480000, "grad_norm": 14.8125, "lr": 1.5e-05, "finish_rate": 0.864, "comp_len": 543.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.0, "frames": {"chat": 221}, "mem_gb": 10.05}
{"step": 5, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.841784818589439, "tokens": 120000, "cumulative_loss_tokens": 600000, "grad_norm": 8.3125, "lr": 1.8e-05, "finish_rate": 0.745, "comp_len": 612.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.3, "frames": {"chat": 196}, "mem_gb": 10.01}
{"step": 6, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.6113010039225221, "tokens": 120000, "cumulative_loss_tokens": 720000, "grad_norm": 8.375, "lr": 2.1e-05, "finish_rate": 0.926, "comp_len": 444.4, "t_data_s": 0.1, "t_rollout_s": 0.0, "t_step_s": 36.6, "frames": {"chat": 270}, "mem_gb": 9.82}
{"step": 7, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.5845785550904771, "tokens": 120000, "cumulative_loss_tokens": 840000, "grad_norm": 4.53125, "lr": 2.4e-05, "finish_rate": 0.815, "comp_len": 555.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.3, "frames": {"chat": 216}, "mem_gb": 9.99}
{"step": 8, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.49846355576987067, "tokens": 120000, "cumulative_loss_tokens": 960000, "grad_norm": 3.390625, "lr": 2.7000000000000002e-05, "finish_rate": 0.775, "comp_len": 600.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 30.9, "frames": {"chat": 200}, "mem_gb": 9.96}
{"step": 9, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.4226150450040897, "tokens": 120000, "cumulative_loss_tokens": 1080000, "grad_norm": 2.609375, "lr": 3e-05, "finish_rate": 0.767, "comp_len": 582.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.8, "frames": {"chat": 206}, "mem_gb": 9.91}
{"step": 10, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.3381338079671065, "tokens": 120000, "cumulative_loss_tokens": 1200000, "grad_norm": 1.7421875, "lr": 3e-05, "finish_rate": 0.902, "comp_len": 512.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.6, "frames": {"chat": 234}, "mem_gb": 9.95}
[eval step 10] sample: "To solve this problem, we need to determine the lengths of the sides of the triangle given the perimeter and the midpoints of its sides. Let's break down the problem into manageable steps:\n\n1. **Under"
{"step": 11, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.34357373327190677, "tokens": 120000, "cumulative_loss_tokens": 1320000, "grad_norm": 1.3828125, "lr": 3e-05, "finish_rate": 0.823, "comp_len": 558.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.3, "frames": {"chat": 215}, "mem_gb": 9.96}
{"step": 12, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.2818106052008768, "tokens": 120000, "cumulative_loss_tokens": 1440000, "grad_norm": 1.1484375, "lr": 3e-05, "finish_rate": 0.922, "comp_len": 470.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.5, "frames": {"chat": 255}, "mem_gb": 9.94}
{"step": 13, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.27427206053423386, "tokens": 120000, "cumulative_loss_tokens": 1560000, "grad_norm": 0.91796875, "lr": 3e-05, "finish_rate": 0.892, "comp_len": 480.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.4, "frames": {"chat": 250}, "mem_gb": 9.82}
{"step": 14, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.2693552000547449, "tokens": 120000, "cumulative_loss_tokens": 1680000, "grad_norm": 0.8359375, "lr": 3e-05, "finish_rate": 0.884, "comp_len": 495.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.1, "frames": {"chat": 242}, "mem_gb": 9.99}
{"step": 15, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.29865186369282504, "tokens": 120000, "cumulative_loss_tokens": 1800000, "grad_norm": 0.8671875, "lr": 3e-05, "finish_rate": 0.729, "comp_len": 603.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.4, "frames": {"chat": 199}, "mem_gb": 10.0}
{"step": 16, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.3619642677543064, "tokens": 120000, "cumulative_loss_tokens": 1920000, "grad_norm": 0.99609375, "lr": 3e-05, "finish_rate": 0.784, "comp_len": 576.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.0, "frames": {"chat": 208}, "mem_gb": 10.04}
{"step": 17, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.25310858338586983, "tokens": 120000, "cumulative_loss_tokens": 2040000, "grad_norm": 0.81640625, "lr": 3e-05, "finish_rate": 0.764, "comp_len": 576.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.4, "frames": {"chat": 208}, "mem_gb": 9.97}
{"step": 18, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.2819262358199805, "tokens": 120000, "cumulative_loss_tokens": 2160000, "grad_norm": 0.80078125, "lr": 3e-05, "finish_rate": 0.732, "comp_len": 574.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.0, "frames": {"chat": 209}, "mem_gb": 10.12}
{"step": 19, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.2152011162099739, "tokens": 120000, "cumulative_loss_tokens": 2280000, "grad_norm": 0.71875, "lr": 3e-05, "finish_rate": 0.855, "comp_len": 510.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.1, "frames": {"chat": 235}, "mem_gb": 9.96}
{"step": 20, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.21725329664709667, "tokens": 120000, "cumulative_loss_tokens": 2400000, "grad_norm": 0.7265625, "lr": 3e-05, "finish_rate": 0.74, "comp_len": 588.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.6, "frames": {"chat": 204}, "mem_gb": 9.95}
[eval step 20] sample: 'To solve this problem, we need to understand the properties of a triangle and its midpoints.\n\n1. **Understanding the Properties:**\n   - The perimeter of a triangle is given by the sum of its sides.\n  '
{"step": 21, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.2732300681939969, "tokens": 120000, "cumulative_loss_tokens": 2520000, "grad_norm": 0.87890625, "lr": 3e-05, "finish_rate": 0.745, "comp_len": 576.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.0, "frames": {"chat": 208}, "mem_gb": 10.01}
{"step": 22, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.20615510911339274, "tokens": 120000, "cumulative_loss_tokens": 2640000, "grad_norm": 0.62109375, "lr": 3e-05, "finish_rate": 0.825, "comp_len": 500.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.7, "frames": {"chat": 240}, "mem_gb": 10.0}
{"step": 23, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.19796610735058784, "tokens": 120000, "cumulative_loss_tokens": 2760000, "grad_norm": 0.62109375, "lr": 3e-05, "finish_rate": 0.89, "comp_len": 487.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.8, "frames": {"chat": 246}, "mem_gb": 9.99}
{"step": 24, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.21532276370519152, "tokens": 120000, "cumulative_loss_tokens": 2880000, "grad_norm": 0.703125, "lr": 3e-05, "finish_rate": 0.909, "comp_len": 493.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.4, "frames": {"chat": 243}, "mem_gb": 9.81}
{"step": 25, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.2322601548684761, "tokens": 120000, "cumulative_loss_tokens": 3000000, "grad_norm": 0.6875, "lr": 3e-05, "finish_rate": 0.745, "comp_len": 576.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.6, "frames": {"chat": 208}, "mem_gb": 10.01}
{"step": 26, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.21518369818814098, "tokens": 120000, "cumulative_loss_tokens": 3120000, "grad_norm": 0.6015625, "lr": 3e-05, "finish_rate": 0.817, "comp_len": 547.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.3, "frames": {"chat": 219}, "mem_gb": 10.0}
{"step": 27, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.23441068366014708, "tokens": 120000, "cumulative_loss_tokens": 3240000, "grad_norm": 0.6953125, "lr": 3e-05, "finish_rate": 0.782, "comp_len": 568.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.5, "frames": {"chat": 211}, "mem_gb": 10.01}
{"step": 28, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.1949662358665218, "tokens": 120000, "cumulative_loss_tokens": 3360000, "grad_norm": 0.62109375, "lr": 3e-05, "finish_rate": 0.862, "comp_len": 517.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.2, "frames": {"chat": 232}, "mem_gb": 9.97}
{"step": 29, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.21759224594825258, "tokens": 120000, "cumulative_loss_tokens": 3480000, "grad_norm": 0.6171875, "lr": 3e-05, "finish_rate": 0.804, "comp_len": 560.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.2, "frames": {"chat": 214}, "mem_gb": 10.01}
{"step": 30, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.18679429283955445, "tokens": 120000, "cumulative_loss_tokens": 3600000, "grad_norm": 0.51953125, "lr": 3e-05, "finish_rate": 0.819, "comp_len": 531.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.9, "frames": {"chat": 226}, "mem_gb": 9.9}
[eval step 30] sample: 'To solve this problem, we need to understand the properties of a triangle and its midpoints.\n\n1. **Understand the Properties:**\n   - The perimeter of a triangle is the sum of its side lengths.\n   - Th'
{"step": 31, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.17500035125799476, "tokens": 120000, "cumulative_loss_tokens": 3720000, "grad_norm": 0.5, "lr": 3e-05, "finish_rate": 0.819, "comp_len": 571.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.3, "frames": {"chat": 210}, "mem_gb": 10.01}
{"step": 32, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.16020464946124702, "tokens": 120000, "cumulative_loss_tokens": 3840000, "grad_norm": 0.486328125, "lr": 3e-05, "finish_rate": 0.858, "comp_len": 550.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.2, "frames": {"chat": 218}, "mem_gb": 9.83}
{"step": 33, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.17072588143019626, "tokens": 120000, "cumulative_loss_tokens": 3960000, "grad_norm": 0.5, "lr": 3e-05, "finish_rate": 0.858, "comp_len": 515.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.8, "frames": {"chat": 233}, "mem_gb": 9.99}
{"step": 34, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.2061376795306181, "tokens": 120000, "cumulative_loss_tokens": 4080000, "grad_norm": 0.6015625, "lr": 3e-05, "finish_rate": 0.786, "comp_len": 558.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.2, "frames": {"chat": 215}, "mem_gb": 10.01}
{"step": 35, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.18651440346712866, "tokens": 120000, "cumulative_loss_tokens": 4200000, "grad_norm": 0.51953125, "lr": 3e-05, "finish_rate": 0.845, "comp_len": 515.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.3, "frames": {"chat": 233}, "mem_gb": 9.99}
{"step": 36, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.16298052088283002, "tokens": 120000, "cumulative_loss_tokens": 4320000, "grad_norm": 0.47265625, "lr": 3e-05, "finish_rate": 0.766, "comp_len": 574.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.0, "frames": {"chat": 209}, "mem_gb": 9.94}
{"step": 37, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.15828031261581926, "tokens": 120000, "cumulative_loss_tokens": 4440000, "grad_norm": 0.4765625, "lr": 3e-05, "finish_rate": 0.908, "comp_len": 458.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.9, "frames": {"chat": 262}, "mem_gb": 9.87}
{"step": 38, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.1595262515474111, "tokens": 120000, "cumulative_loss_tokens": 4560000, "grad_norm": 0.466796875, "lr": 3e-05, "finish_rate": 0.9, "comp_len": 481.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.1, "frames": {"chat": 249}, "mem_gb": 9.96}
{"step": 39, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.19758843879966687, "tokens": 120000, "cumulative_loss_tokens": 4680000, "grad_norm": 0.56640625, "lr": 3e-05, "finish_rate": 0.819, "comp_len": 528.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.5, "frames": {"chat": 227}, "mem_gb": 10.0}
{"step": 40, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.14780844743441168, "tokens": 120000, "cumulative_loss_tokens": 4800000, "grad_norm": 0.4375, "lr": 3e-05, "finish_rate": 0.814, "comp_len": 543.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.5, "frames": {"chat": 221}, "mem_gb": 9.99}
[eval step 40] sample: "To solve this problem, we need to understand the properties of the midpoints of a triangle's sides and how they affect the perimeter.\n\n1. **Midpoints of Sides:**\n   - The midpoint of a side of length "
{"step": 41, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.16207660000277682, "tokens": 120000, "cumulative_loss_tokens": 4920000, "grad_norm": 0.458984375, "lr": 3e-05, "finish_rate": 0.859, "comp_len": 512.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.5, "frames": {"chat": 234}, "mem_gb": 10.01}
{"step": 42, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.13810028650884826, "tokens": 120000, "cumulative_loss_tokens": 5040000, "grad_norm": 0.42578125, "lr": 3e-05, "finish_rate": 0.817, "comp_len": 563.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.3, "frames": {"chat": 213}, "mem_gb": 9.96}
{"step": 43, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.13506372714247555, "tokens": 120000, "cumulative_loss_tokens": 5160000, "grad_norm": 0.40234375, "lr": 3e-05, "finish_rate": 0.836, "comp_len": 563.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.5, "frames": {"chat": 213}, "mem_gb": 9.89}
{"step": 44, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.14545865754342327, "tokens": 120000, "cumulative_loss_tokens": 5280000, "grad_norm": 0.43359375, "lr": 3e-05, "finish_rate": 0.906, "comp_len": 512.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.9, "frames": {"chat": 234}, "mem_gb": 9.92}
{"step": 45, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.14481995174276333, "tokens": 120000, "cumulative_loss_tokens": 5400000, "grad_norm": 0.423828125, "lr": 3e-05, "finish_rate": 0.793, "comp_len": 540.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.8, "frames": {"chat": 222}, "mem_gb": 9.99}
{"step": 46, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.17587772664371878, "tokens": 120000, "cumulative_loss_tokens": 5520000, "grad_norm": 0.498046875, "lr": 3e-05, "finish_rate": 0.806, "comp_len": 528.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 36.1, "frames": {"chat": 227}, "mem_gb": 10.01}
{"step": 47, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.1550076254642258, "tokens": 120000, "cumulative_loss_tokens": 5640000, "grad_norm": 0.427734375, "lr": 3e-05, "finish_rate": 0.835, "comp_len": 550.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.4, "frames": {"chat": 218}, "mem_gb": 10.04}
{"step": 48, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.17269171449063966, "tokens": 120000, "cumulative_loss_tokens": 5760000, "grad_norm": 0.5, "lr": 3e-05, "finish_rate": 0.798, "comp_len": 538.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.1, "frames": {"chat": 223}, "mem_gb": 10.01}
{"step": 49, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.17547114912066608, "tokens": 120000, "cumulative_loss_tokens": 5880000, "grad_norm": 0.4609375, "lr": 3e-05, "finish_rate": 0.772, "comp_len": 582.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.1, "frames": {"chat": 206}, "mem_gb": 10.01}
{"step": 50, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.1537847354217743, "tokens": 120000, "cumulative_loss_tokens": 6000000, "grad_norm": 0.490234375, "lr": 3e-05, "finish_rate": 0.784, "comp_len": 563.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.2, "frames": {"chat": 213}, "mem_gb": 9.92}
[eval step 50] sample: "To solve this problem, we need to understand the geometric properties involved. Here's a step-by-step breakdown:\n\n1. **Understand the Problem:**\n   - The perimeter of the original triangle is given as"
checkpoint snapshot queued -> outputs/healed/grid_math/glean_keep25_s1226/step0050
{"step": 51, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.18943238889773686, "tokens": 120000, "cumulative_loss_tokens": 6120000, "grad_norm": 0.50390625, "lr": 3e-05, "finish_rate": 0.843, "comp_len": 538.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.9, "frames": {"chat": 223}, "mem_gb": 9.86}
{"step": 52, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.16693324066617837, "tokens": 120000, "cumulative_loss_tokens": 6240000, "grad_norm": 0.4609375, "lr": 3e-05, "finish_rate": 0.828, "comp_len": 528.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.6, "frames": {"chat": 227}, "mem_gb": 9.97}
{"step": 53, "epoch": 0, "training_mode": "off-policy", "forward_topk_kl": 0.14277432735928644, "tokens": 120000, "cumulative_loss_tokens": 6360000, "grad_norm": 0.4453125, "lr": 3e-05, "finish_rate": 0.889, "comp_len": 474.3, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.6, "frames": {"chat": 253}, "mem_gb": 10.0}
{"step": 54, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.1437762385570444, "tokens": 120000, "cumulative_loss_tokens": 6480000, "grad_norm": 0.83984375, "lr": 3e-05, "finish_rate": 0.792, "comp_len": 555.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.9, "frames": {"chat": 216}, "mem_gb": 10.01}
{"step": 55, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.1300227014787495, "tokens": 120000, "cumulative_loss_tokens": 6600000, "grad_norm": 0.412109375, "lr": 3e-05, "finish_rate": 0.766, "comp_len": 585.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.7, "frames": {"chat": 205}, "mem_gb": 9.98}
{"step": 56, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.15664660246831674, "tokens": 120000, "cumulative_loss_tokens": 6720000, "grad_norm": 0.46875, "lr": 3e-05, "finish_rate": 0.729, "comp_len": 579.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.8, "frames": {"chat": 207}, "mem_gb": 10.06}
{"step": 57, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.1445078781637363, "tokens": 120000, "cumulative_loss_tokens": 6840000, "grad_norm": 0.4140625, "lr": 3e-05, "finish_rate": 0.814, "comp_len": 558.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.0, "frames": {"chat": 215}, "mem_gb": 9.98}
{"step": 58, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10567276090113446, "tokens": 120000, "cumulative_loss_tokens": 6960000, "grad_norm": 0.373046875, "lr": 3e-05, "finish_rate": 0.86, "comp_len": 526.3, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.9, "frames": {"chat": 228}, "mem_gb": 10.0}
{"step": 59, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.13750520292868218, "tokens": 120000, "cumulative_loss_tokens": 7080000, "grad_norm": 0.47265625, "lr": 3e-05, "finish_rate": 0.747, "comp_len": 543.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.3, "frames": {"chat": 221}, "mem_gb": 10.04}
{"step": 60, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10064160097377996, "tokens": 120000, "cumulative_loss_tokens": 7200000, "grad_norm": 0.380859375, "lr": 3e-05, "finish_rate": 0.882, "comp_len": 472.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.1, "frames": {"chat": 254}, "mem_gb": 9.84}
[eval step 60] sample: "To solve this problem, we need to understand the geometric properties involved. Here's a step-by-step breakdown:\n\n1. **Understand the Problem:**\n   - The perimeter of the original triangle is given as"
{"step": 61, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.09463851169059052, "tokens": 120000, "cumulative_loss_tokens": 7320000, "grad_norm": 0.375, "lr": 3e-05, "finish_rate": 0.843, "comp_len": 571.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.9, "frames": {"chat": 210}, "mem_gb": 9.97}
{"step": 62, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10595882567154864, "tokens": 120000, "cumulative_loss_tokens": 7440000, "grad_norm": 0.345703125, "lr": 3e-05, "finish_rate": 0.827, "comp_len": 531.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.2, "frames": {"chat": 226}, "mem_gb": 9.92}
{"step": 63, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.12365473369524503, "tokens": 120000, "cumulative_loss_tokens": 7560000, "grad_norm": 0.38671875, "lr": 3e-05, "finish_rate": 0.802, "comp_len": 566.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.3, "frames": {"chat": 212}, "mem_gb": 9.99}
{"step": 64, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.12891767050419004, "tokens": 120000, "cumulative_loss_tokens": 7680000, "grad_norm": 0.365234375, "lr": 3e-05, "finish_rate": 0.754, "comp_len": 568.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.7, "frames": {"chat": 211}, "mem_gb": 9.93}
{"step": 65, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.13171432805840547, "tokens": 120000, "cumulative_loss_tokens": 7800000, "grad_norm": 0.41015625, "lr": 3e-05, "finish_rate": 0.776, "comp_len": 612.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 30.4, "frames": {"chat": 196}, "mem_gb": 9.97}
{"step": 66, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.11064968370975306, "tokens": 120000, "cumulative_loss_tokens": 7920000, "grad_norm": 0.38671875, "lr": 3e-05, "finish_rate": 0.811, "comp_len": 566.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.0, "frames": {"chat": 212}, "mem_gb": 10.0}
{"step": 67, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.1052026781039002, "tokens": 120000, "cumulative_loss_tokens": 8040000, "grad_norm": 0.35546875, "lr": 3e-05, "finish_rate": 0.877, "comp_len": 491.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.8, "frames": {"chat": 244}, "mem_gb": 9.91}
{"step": 68, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10334055133834481, "tokens": 120000, "cumulative_loss_tokens": 8160000, "grad_norm": 0.326171875, "lr": 3e-05, "finish_rate": 0.838, "comp_len": 540.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.5, "frames": {"chat": 222}, "mem_gb": 9.95}
{"step": 69, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.11475445196203267, "tokens": 120000, "cumulative_loss_tokens": 8280000, "grad_norm": 0.36328125, "lr": 3e-05, "finish_rate": 0.78, "comp_len": 550.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.3, "frames": {"chat": 218}, "mem_gb": 10.0}
{"step": 70, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.12215668223105992, "tokens": 120000, "cumulative_loss_tokens": 8400000, "grad_norm": 0.380859375, "lr": 3e-05, "finish_rate": 0.913, "comp_len": 476.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 35.2, "frames": {"chat": 252}, "mem_gb": 9.88}
[eval step 70] sample: 'To solve this problem, we need to understand the geometric properties of the triangle formed by the midpoints of the sides of the original triangle.\n\n1. **Identify the Original Triangle:**\n   Let the '
{"step": 71, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.13357413773555307, "tokens": 120000, "cumulative_loss_tokens": 8520000, "grad_norm": 0.431640625, "lr": 3e-05, "finish_rate": 0.728, "comp_len": 594.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.7, "frames": {"chat": 202}, "mem_gb": 10.05}
{"step": 72, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.15701979057999949, "tokens": 120000, "cumulative_loss_tokens": 8640000, "grad_norm": 0.5, "lr": 3e-05, "finish_rate": 0.835, "comp_len": 506.3, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.4, "frames": {"chat": 237}, "mem_gb": 10.0}
{"step": 73, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.11951561769278099, "tokens": 120000, "cumulative_loss_tokens": 8760000, "grad_norm": 0.359375, "lr": 3e-05, "finish_rate": 0.868, "comp_len": 512.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.5, "frames": {"chat": 234}, "mem_gb": 9.99}
{"step": 74, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.09150583964148536, "tokens": 120000, "cumulative_loss_tokens": 8880000, "grad_norm": 0.33203125, "lr": 3e-05, "finish_rate": 0.809, "comp_len": 558.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.5, "frames": {"chat": 215}, "mem_gb": 10.0}
{"step": 75, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.08957106544353689, "tokens": 120000, "cumulative_loss_tokens": 9000000, "grad_norm": 0.333984375, "lr": 3e-05, "finish_rate": 0.876, "comp_len": 512.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.9, "frames": {"chat": 234}, "mem_gb": 9.93}
{"step": 76, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.08833280240970974, "tokens": 120000, "cumulative_loss_tokens": 9120000, "grad_norm": 0.3125, "lr": 3e-05, "finish_rate": 0.801, "comp_len": 555.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.7, "frames": {"chat": 216}, "mem_gb": 9.99}
{"step": 77, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.1060180736657232, "tokens": 120000, "cumulative_loss_tokens": 9240000, "grad_norm": 0.34375, "lr": 3e-05, "finish_rate": 0.805, "comp_len": 571.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.3, "frames": {"chat": 210}, "mem_gb": 9.95}
{"step": 78, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10358950277809054, "tokens": 120000, "cumulative_loss_tokens": 9360000, "grad_norm": 0.34375, "lr": 3e-05, "finish_rate": 0.719, "comp_len": 603.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.4, "frames": {"chat": 199}, "mem_gb": 10.0}
{"step": 79, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.09346498899217695, "tokens": 120000, "cumulative_loss_tokens": 9480000, "grad_norm": 0.318359375, "lr": 3e-05, "finish_rate": 0.824, "comp_len": 571.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.0, "frames": {"chat": 210}, "mem_gb": 10.01}
{"step": 80, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10706426500252758, "tokens": 120000, "cumulative_loss_tokens": 9600000, "grad_norm": 0.40625, "lr": 3e-05, "finish_rate": 0.902, "comp_len": 533.3, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.2, "frames": {"chat": 225}, "mem_gb": 9.95}
[eval step 80] sample: 'To solve this problem, we need to understand the geometric properties of the triangle and its midpoints.\n\n1. **Understand the Problem:**\n   - The perimeter of the original triangle is given as 28.\n   '
{"step": 81, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10597167926300317, "tokens": 120000, "cumulative_loss_tokens": 9720000, "grad_norm": 0.37109375, "lr": 3e-05, "finish_rate": 0.913, "comp_len": 474.3, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.2, "frames": {"chat": 253}, "mem_gb": 9.85}
{"step": 82, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10689509418696785, "tokens": 120000, "cumulative_loss_tokens": 9840000, "grad_norm": 0.359375, "lr": 3e-05, "finish_rate": 0.903, "comp_len": 485.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.6, "frames": {"chat": 247}, "mem_gb": 9.97}
{"step": 83, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10978524073281636, "tokens": 120000, "cumulative_loss_tokens": 9960000, "grad_norm": 0.3671875, "lr": 3e-05, "finish_rate": 0.836, "comp_len": 504.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.4, "frames": {"chat": 238}, "mem_gb": 9.97}
{"step": 84, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.11432554268656919, "tokens": 120000, "cumulative_loss_tokens": 10080000, "grad_norm": 0.380859375, "lr": 3e-05, "finish_rate": 0.86, "comp_len": 510.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.8, "frames": {"chat": 235}, "mem_gb": 10.0}
{"step": 85, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.11262670917728294, "tokens": 120000, "cumulative_loss_tokens": 10200000, "grad_norm": 0.392578125, "lr": 3e-05, "finish_rate": 0.805, "comp_len": 558.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.8, "frames": {"chat": 215}, "mem_gb": 9.97}
{"step": 86, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.11482771853323405, "tokens": 120000, "cumulative_loss_tokens": 10320000, "grad_norm": 0.443359375, "lr": 3e-05, "finish_rate": 0.925, "comp_len": 447.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.2, "frames": {"chat": 268}, "mem_gb": 9.97}
{"step": 87, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10706278533112878, "tokens": 120000, "cumulative_loss_tokens": 10440000, "grad_norm": 0.3515625, "lr": 3e-05, "finish_rate": 0.825, "comp_len": 526.3, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.3, "frames": {"chat": 228}, "mem_gb": 10.0}
{"step": 88, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.11632729308865965, "tokens": 120000, "cumulative_loss_tokens": 10560000, "grad_norm": 0.38671875, "lr": 3e-05, "finish_rate": 0.881, "comp_len": 476.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.1, "frames": {"chat": 252}, "mem_gb": 9.93}
{"step": 89, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.09904835979019602, "tokens": 120000, "cumulative_loss_tokens": 10680000, "grad_norm": 0.365234375, "lr": 3e-05, "finish_rate": 0.821, "comp_len": 538.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.7, "frames": {"chat": 223}, "mem_gb": 10.01}
{"step": 90, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.15139277683890734, "tokens": 120000, "cumulative_loss_tokens": 10800000, "grad_norm": 0.45703125, "lr": 3e-05, "finish_rate": 0.805, "comp_len": 531.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.6, "frames": {"chat": 226}, "mem_gb": 10.0}
[eval step 90] sample: 'To solve this problem, we need to understand the geometric properties of the triangle formed by the midpoints of the sides of a triangle with a given perimeter.\n\n### Steps to Solve:\n\n1. **Understand t'
{"step": 91, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.12696473920823384, "tokens": 120000, "cumulative_loss_tokens": 10920000, "grad_norm": 0.396484375, "lr": 3e-05, "finish_rate": 0.731, "comp_len": 576.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 35.0, "frames": {"chat": 208}, "mem_gb": 10.05}
{"step": 92, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.0981225883629794, "tokens": 120000, "cumulative_loss_tokens": 11040000, "grad_norm": 0.341796875, "lr": 3e-05, "finish_rate": 0.883, "comp_len": 500.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.6, "frames": {"chat": 240}, "mem_gb": 9.93}
{"step": 93, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.11373697855863721, "tokens": 120000, "cumulative_loss_tokens": 11160000, "grad_norm": 0.390625, "lr": 3e-05, "finish_rate": 0.842, "comp_len": 540.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.9, "frames": {"chat": 222}, "mem_gb": 9.93}
{"step": 94, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.09599229777337362, "tokens": 120000, "cumulative_loss_tokens": 11280000, "grad_norm": 0.3359375, "lr": 3e-05, "finish_rate": 0.881, "comp_len": 508.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.7, "frames": {"chat": 236}, "mem_gb": 10.0}
{"step": 95, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.0905342026155442, "tokens": 120000, "cumulative_loss_tokens": 11400000, "grad_norm": 0.330078125, "lr": 3e-05, "finish_rate": 0.834, "comp_len": 553.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.9, "frames": {"chat": 217}, "mem_gb": 9.96}
{"step": 96, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10233824915119136, "tokens": 120000, "cumulative_loss_tokens": 11520000, "grad_norm": 0.392578125, "lr": 3e-05, "finish_rate": 0.921, "comp_len": 476.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.8, "frames": {"chat": 252}, "mem_gb": 9.88}
{"step": 97, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.09226011450669418, "tokens": 120000, "cumulative_loss_tokens": 11640000, "grad_norm": 0.3359375, "lr": 3e-05, "finish_rate": 0.847, "comp_len": 540.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.4, "frames": {"chat": 222}, "mem_gb": 9.99}
{"step": 98, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.1017231995769466, "tokens": 120000, "cumulative_loss_tokens": 11760000, "grad_norm": 0.3671875, "lr": 3e-05, "finish_rate": 0.901, "comp_len": 495.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.4, "frames": {"chat": 242}, "mem_gb": 9.87}
{"step": 99, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.13698266774720202, "tokens": 120000, "cumulative_loss_tokens": 11880000, "grad_norm": 0.396484375, "lr": 3e-05, "finish_rate": 0.804, "comp_len": 547.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.2, "frames": {"chat": 219}, "mem_gb": 9.93}
{"step": 100, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.09786459891942019, "tokens": 120000, "cumulative_loss_tokens": 12000000, "grad_norm": 0.333984375, "lr": 3e-05, "finish_rate": 0.879, "comp_len": 538.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.9, "frames": {"chat": 223}, "mem_gb": 9.94}
[eval step 100] sample: "To solve this problem, we need to understand the geometric properties involved. Here's a step-by-step breakdown:\n\n1. **Understand the Problem:**\n   - The perimeter of the original triangle is 28.\n   -"
checkpoint snapshot queued -> outputs/healed/grid_math/glean_keep25_s1226/step0100
{"step": 101, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.11257973559337357, "tokens": 120000, "cumulative_loss_tokens": 12120000, "grad_norm": 0.37890625, "lr": 3e-05, "finish_rate": 0.858, "comp_len": 517.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.4, "frames": {"chat": 232}, "mem_gb": 9.95}
{"step": 102, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.11817761343022187, "tokens": 120000, "cumulative_loss_tokens": 12240000, "grad_norm": 0.416015625, "lr": 3e-05, "finish_rate": 0.832, "comp_len": 545.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.5, "frames": {"chat": 220}, "mem_gb": 10.0}
{"step": 103, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.1162180175398166, "tokens": 120000, "cumulative_loss_tokens": 12360000, "grad_norm": 0.37109375, "lr": 3e-05, "finish_rate": 0.776, "comp_len": 571.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.9, "frames": {"chat": 210}, "mem_gb": 10.04}
{"step": 104, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.10464485003838005, "tokens": 120000, "cumulative_loss_tokens": 12480000, "grad_norm": 0.33984375, "lr": 3e-05, "finish_rate": 0.81, "comp_len": 531.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.0, "frames": {"chat": 226}, "mem_gb": 9.97}
{"step": 105, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.09739518145142743, "tokens": 120000, "cumulative_loss_tokens": 12600000, "grad_norm": 0.34375, "lr": 3e-05, "finish_rate": 0.741, "comp_len": 566.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.9, "frames": {"chat": 212}, "mem_gb": 10.0}
{"step": 106, "epoch": 1, "training_mode": "off-policy", "forward_topk_kl": 0.09368765245955438, "tokens": 120000, "cumulative_loss_tokens": 12720000, "grad_norm": 0.58203125, "lr": 3e-05, "finish_rate": 0.839, "comp_len": 508.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.3, "frames": {"chat": 236}, "mem_gb": 10.01}
{"step": 107, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07762365770225103, "tokens": 120000, "cumulative_loss_tokens": 12840000, "grad_norm": 0.349609375, "lr": 3e-05, "finish_rate": 0.928, "comp_len": 454.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.0, "frames": {"chat": 264}, "mem_gb": 9.88}
{"step": 108, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.1221225647800602, "tokens": 120000, "cumulative_loss_tokens": 12960000, "grad_norm": 0.52734375, "lr": 3e-05, "finish_rate": 0.834, "comp_len": 524.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.6, "frames": {"chat": 229}, "mem_gb": 9.98}
{"step": 109, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.0774544737749733, "tokens": 120000, "cumulative_loss_tokens": 13080000, "grad_norm": 0.29296875, "lr": 3e-05, "finish_rate": 0.903, "comp_len": 465.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.3, "frames": {"chat": 258}, "mem_gb": 9.85}
{"step": 110, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.11255108813242987, "tokens": 120000, "cumulative_loss_tokens": 13200000, "grad_norm": 0.3984375, "lr": 3e-05, "finish_rate": 0.755, "comp_len": 576.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.7, "frames": {"chat": 208}, "mem_gb": 10.01}
[eval step 110] sample: 'To solve this problem, we need to understand the geometric properties of the triangle formed by the midpoints of the sides of the original triangle.\n\n1. **Midpoints of a Triangle:**\n   - The midpoint '
{"step": 111, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.0868233092402108, "tokens": 120000, "cumulative_loss_tokens": 13320000, "grad_norm": 0.341796875, "lr": 3e-05, "finish_rate": 0.88, "comp_len": 481.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.7, "frames": {"chat": 249}, "mem_gb": 9.93}
{"step": 112, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07358991829988858, "tokens": 120000, "cumulative_loss_tokens": 13440000, "grad_norm": 0.296875, "lr": 3e-05, "finish_rate": 0.845, "comp_len": 545.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.5, "frames": {"chat": 220}, "mem_gb": 9.99}
{"step": 113, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.08318625353276729, "tokens": 120000, "cumulative_loss_tokens": 13560000, "grad_norm": 0.31640625, "lr": 3e-05, "finish_rate": 0.834, "comp_len": 538.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.8, "frames": {"chat": 223}, "mem_gb": 9.99}
{"step": 114, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07259801399639497, "tokens": 120000, "cumulative_loss_tokens": 13680000, "grad_norm": 0.32421875, "lr": 3e-05, "finish_rate": 0.833, "comp_len": 543.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.3, "frames": {"chat": 221}, "mem_gb": 10.0}
{"step": 115, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07171069395930196, "tokens": 120000, "cumulative_loss_tokens": 13800000, "grad_norm": 0.2890625, "lr": 3e-05, "finish_rate": 0.9, "comp_len": 521.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.6, "frames": {"chat": 230}, "mem_gb": 9.9}
{"step": 116, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.08987705074110999, "tokens": 120000, "cumulative_loss_tokens": 13920000, "grad_norm": 0.3125, "lr": 3e-05, "finish_rate": 0.776, "comp_len": 560.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.6, "frames": {"chat": 214}, "mem_gb": 9.97}
{"step": 117, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.11477152344050506, "tokens": 120000, "cumulative_loss_tokens": 14040000, "grad_norm": 0.35546875, "lr": 3e-05, "finish_rate": 0.766, "comp_len": 560.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.2, "frames": {"chat": 214}, "mem_gb": 9.99}
{"step": 118, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.08955526669428994, "tokens": 120000, "cumulative_loss_tokens": 14160000, "grad_norm": 0.330078125, "lr": 3e-05, "finish_rate": 0.786, "comp_len": 571.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.8, "frames": {"chat": 210}, "mem_gb": 10.04}
{"step": 119, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.09731406231168657, "tokens": 120000, "cumulative_loss_tokens": 14280000, "grad_norm": 0.3359375, "lr": 3e-05, "finish_rate": 0.776, "comp_len": 560.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.6, "frames": {"chat": 214}, "mem_gb": 10.0}
{"step": 120, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.0877075838214097, "tokens": 120000, "cumulative_loss_tokens": 14400000, "grad_norm": 0.302734375, "lr": 3e-05, "finish_rate": 0.791, "comp_len": 558.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.5, "frames": {"chat": 215}, "mem_gb": 9.95}
[eval step 120] sample: "To solve this problem, we need to understand the geometric properties involved. Here's a step-by-step breakdown:\n\n1. **Understand the Problem:**\n   - The perimeter of the triangle is 28.\n   - The midp"
{"step": 121, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.09859609124142056, "tokens": 120000, "cumulative_loss_tokens": 14520000, "grad_norm": 0.349609375, "lr": 3e-05, "finish_rate": 0.721, "comp_len": 576.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.8, "frames": {"chat": 208}, "mem_gb": 9.99}
{"step": 122, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07579588066336389, "tokens": 120000, "cumulative_loss_tokens": 14640000, "grad_norm": 0.291015625, "lr": 3e-05, "finish_rate": 0.789, "comp_len": 550.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.7, "frames": {"chat": 218}, "mem_gb": 9.88}
{"step": 123, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.06975389175747211, "tokens": 120000, "cumulative_loss_tokens": 14760000, "grad_norm": 0.265625, "lr": 3e-05, "finish_rate": 0.876, "comp_len": 515.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.9, "frames": {"chat": 233}, "mem_gb": 9.9}
{"step": 124, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.06521768421179926, "tokens": 120000, "cumulative_loss_tokens": 14880000, "grad_norm": 0.267578125, "lr": 3e-05, "finish_rate": 0.861, "comp_len": 519.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.1, "frames": {"chat": 231}, "mem_gb": 9.94}
{"step": 125, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.08123537786286324, "tokens": 120000, "cumulative_loss_tokens": 15000000, "grad_norm": 0.306640625, "lr": 3e-05, "finish_rate": 0.868, "comp_len": 510.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.2, "frames": {"chat": 235}, "mem_gb": 10.13}
{"step": 126, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07916606919132173, "tokens": 120000, "cumulative_loss_tokens": 15120000, "grad_norm": 0.27734375, "lr": 3e-05, "finish_rate": 0.843, "comp_len": 555.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.0, "frames": {"chat": 216}, "mem_gb": 9.99}
{"step": 127, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.0735622343561612, "tokens": 120000, "cumulative_loss_tokens": 15240000, "grad_norm": 0.28125, "lr": 3e-05, "finish_rate": 0.831, "comp_len": 506.3, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.4, "frames": {"chat": 237}, "mem_gb": 10.01}
{"step": 128, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.10053933701403439, "tokens": 120000, "cumulative_loss_tokens": 15360000, "grad_norm": 0.333984375, "lr": 3e-05, "finish_rate": 0.734, "comp_len": 591.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.4, "frames": {"chat": 203}, "mem_gb": 10.01}
{"step": 129, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07754695314977629, "tokens": 120000, "cumulative_loss_tokens": 15480000, "grad_norm": 0.310546875, "lr": 3e-05, "finish_rate": 0.873, "comp_len": 508.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.8, "frames": {"chat": 236}, "mem_gb": 10.04}
{"step": 130, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07985588553401952, "tokens": 120000, "cumulative_loss_tokens": 15600000, "grad_norm": 0.2890625, "lr": 3e-05, "finish_rate": 0.734, "comp_len": 560.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.2, "frames": {"chat": 214}, "mem_gb": 10.01}
[eval step 130] sample: 'To solve this problem, we need to understand the geometric properties of the triangle formed by the midpoints of the sides of a given triangle.\n\n1. **Understand the Midpoints:**\n   - The midpoint of a'
{"step": 131, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.0826984562702477, "tokens": 120000, "cumulative_loss_tokens": 15720000, "grad_norm": 0.302734375, "lr": 3e-05, "finish_rate": 0.78, "comp_len": 574.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.1, "frames": {"chat": 209}, "mem_gb": 10.0}
{"step": 132, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07694104558015243, "tokens": 120000, "cumulative_loss_tokens": 15840000, "grad_norm": 0.306640625, "lr": 3e-05, "finish_rate": 0.906, "comp_len": 468.8, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.9, "frames": {"chat": 256}, "mem_gb": 10.0}
{"step": 133, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07371389015351111, "tokens": 120000, "cumulative_loss_tokens": 15960000, "grad_norm": 0.29296875, "lr": 3e-05, "finish_rate": 0.878, "comp_len": 521.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.3, "frames": {"chat": 230}, "mem_gb": 9.87}
{"step": 134, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07886174646293123, "tokens": 120000, "cumulative_loss_tokens": 16080000, "grad_norm": 0.29296875, "lr": 3e-05, "finish_rate": 0.822, "comp_len": 521.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.6, "frames": {"chat": 230}, "mem_gb": 10.06}
{"step": 135, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.09088345281931882, "tokens": 120000, "cumulative_loss_tokens": 16200000, "grad_norm": 0.322265625, "lr": 3e-05, "finish_rate": 0.881, "comp_len": 528.6, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.8, "frames": {"chat": 227}, "mem_gb": 9.95}
{"step": 136, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.09291109785580387, "tokens": 120000, "cumulative_loss_tokens": 16320000, "grad_norm": 0.31640625, "lr": 3e-05, "finish_rate": 0.755, "comp_len": 576.9, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.8, "frames": {"chat": 208}, "mem_gb": 10.01}
{"step": 137, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.09028420611228793, "tokens": 120000, "cumulative_loss_tokens": 16440000, "grad_norm": 0.302734375, "lr": 3e-05, "finish_rate": 0.699, "comp_len": 582.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.2, "frames": {"chat": 206}, "mem_gb": 10.03}
{"step": 138, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.08365592800673718, "tokens": 120000, "cumulative_loss_tokens": 16560000, "grad_norm": 0.314453125, "lr": 3e-05, "finish_rate": 0.82, "comp_len": 526.3, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.7, "frames": {"chat": 228}, "mem_gb": 9.9}
{"step": 139, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.08592219653519181, "tokens": 120000, "cumulative_loss_tokens": 16680000, "grad_norm": 0.306640625, "lr": 3e-05, "finish_rate": 0.835, "comp_len": 535.7, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.7, "frames": {"chat": 224}, "mem_gb": 10.0}
{"step": 140, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07787073799719413, "tokens": 120000, "cumulative_loss_tokens": 16800000, "grad_norm": 0.2890625, "lr": 3e-05, "finish_rate": 0.66, "comp_len": 600.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.4, "frames": {"chat": 200}, "mem_gb": 10.03}
[eval step 140] sample: 'To solve this problem, we need to understand the geometric properties of the triangle formed by the midpoints of the sides of the original triangle.\n\n1. **Understand the Midpoints:**\n   The midpoints '
{"step": 141, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07876405616238868, "tokens": 120000, "cumulative_loss_tokens": 16920000, "grad_norm": 0.296875, "lr": 3e-05, "finish_rate": 0.714, "comp_len": 612.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 30.8, "frames": {"chat": 196}, "mem_gb": 10.01}
{"step": 142, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.0733729308873415, "tokens": 120000, "cumulative_loss_tokens": 17040000, "grad_norm": 0.283203125, "lr": 3e-05, "finish_rate": 0.834, "comp_len": 538.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.4, "frames": {"chat": 223}, "mem_gb": 9.99}
{"step": 143, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07316665141967435, "tokens": 120000, "cumulative_loss_tokens": 17160000, "grad_norm": 0.291015625, "lr": 3e-05, "finish_rate": 0.869, "comp_len": 563.4, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 31.4, "frames": {"chat": 213}, "mem_gb": 9.89}
{"step": 144, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.06799636307867865, "tokens": 120000, "cumulative_loss_tokens": 17280000, "grad_norm": 0.287109375, "lr": 3e-05, "finish_rate": 0.879, "comp_len": 517.2, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 33.1, "frames": {"chat": 232}, "mem_gb": 9.93}
{"step": 145, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.06842424086419245, "tokens": 120000, "cumulative_loss_tokens": 17400000, "grad_norm": 0.27734375, "lr": 3e-05, "finish_rate": 0.861, "comp_len": 538.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.5, "frames": {"chat": 223}, "mem_gb": 9.93}
{"step": 146, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.07442820003066833, "tokens": 120000, "cumulative_loss_tokens": 17520000, "grad_norm": 0.29296875, "lr": 3e-05, "finish_rate": 0.85, "comp_len": 515.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.9, "frames": {"chat": 233}, "mem_gb": 10.02}
{"step": 147, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.08168693628491212, "tokens": 120000, "cumulative_loss_tokens": 17640000, "grad_norm": 0.287109375, "lr": 3e-05, "finish_rate": 0.816, "comp_len": 553.0, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.1, "frames": {"chat": 217}, "mem_gb": 10.01}
{"step": 148, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.11692778237918391, "tokens": 120000, "cumulative_loss_tokens": 17760000, "grad_norm": 0.36328125, "lr": 3e-05, "finish_rate": 0.752, "comp_len": 594.1, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.6, "frames": {"chat": 202}, "mem_gb": 10.08}
{"step": 149, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.0718278338014769, "tokens": 120000, "cumulative_loss_tokens": 17880000, "grad_norm": 0.283203125, "lr": 3e-05, "finish_rate": 0.858, "comp_len": 474.3, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 34.2, "frames": {"chat": 253}, "mem_gb": 9.93}
{"step": 150, "epoch": 2, "training_mode": "off-policy", "forward_topk_kl": 0.06792547624123593, "tokens": 120000, "cumulative_loss_tokens": 18000000, "grad_norm": 0.263671875, "lr": 3e-05, "finish_rate": 0.879, "comp_len": 519.5, "t_data_s": 0.0, "t_rollout_s": 0.0, "t_step_s": 32.7, "frames": {"chat": 231}, "mem_gb": 9.94}
[eval step 150] sample: 'To solve this problem, we need to understand the geometric properties of the triangle formed by the midpoints of the sides of the original triangle.\n\n1. **Midpoints of a Triangle:**\n   The midpoint of'
checkpoint snapshot queued -> outputs/healed/grid_math/glean_keep25_s1226/step0150
wandb: updating run metadata
wandb: uploading output.log; uploading wandb-summary.json; uploading config.yaml
wandb: uploading summary, console lines 170-170
wandb: 
wandb: Run history:
wandb:               comp_len █▆▂█▃▂▆▅▅▆▁▅▅▅▄▆▄▆▃▄▆▂▃▄▂▄▅▂▃▅▆▃▁▆▅▆▃▅▃▆
wandb: cumulative_loss_tokens ▁▁▂▂▂▂▂▂▂▂▃▃▃▃▃▄▄▄▄▄▅▅▅▅▅▅▅▆▆▆▆▆▇▇▇▇▇███
wandb:                  epoch ▁▁▁▁▁▁▁▁▁▁▁▁▁▅▅▅▅▅▅▅▅▅▅▅▅███████████████
wandb:            finish_rate ▇▇▃▆█▃▆▃▅▅▅▄▆▆▄▃▃▃▆▆▇▅▃▆▇▇▇▅▃█▇▆▄▆▆▇▂▅▁▆
wandb:        forward_topk_kl ██▆▅▄▂▂▂▂▂▂▂▂▂▂▂▂▂▂▁▁▁▁▁▁▁▁▁▂▁▁▁▁▁▁▁▁▁▁▁
wandb:              grad_norm █▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
wandb:                     lr ▁███████████████████████████████████████
wandb:                 mem_gb ▆▅▂▃▄▃▅▅▄▄▄▅▅▆▃▄▅▆▄▆▅▁▅▂▄▆▁▅▄▅▃█▄▅▅▅▅▂▅▃
wandb:                   step ▁▁▁▁▁▂▂▂▂▂▂▂▃▃▃▃▃▃▃▄▄▄▄▅▅▅▅▆▆▆▆▆▇▇▇▇▇▇██
wandb:               t_data_s ▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁▁
wandb:                     +3 ...
wandb: 
wandb: Run summary:
wandb:               comp_len 519.5
wandb: cumulative_loss_tokens 18000000
wandb:                  epoch 2
wandb:            finish_rate 0.879
wandb:        forward_topk_kl 0.06793
wandb:              grad_norm 0.26367
wandb:                     lr 3e-05
wandb:                 mem_gb 9.94
wandb:                   step 150
wandb:               t_data_s 0
wandb:                     +4 ...
wandb: 
wandb: 🚀 View run glean-math-keep25-s1226 at: https://wandb.ai/hbfreed/glean-grid/runs/vhgtf0ej
wandb: ⭐️ View project at: https://wandb.ai/hbfreed/glean-grid
wandb: Synced 5 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other file(s)
wandb: Find logs at: outputs/healed/grid_math/glean_keep25_s1226/wandb/run-20260716_034735-vhgtf0ej/logs
{
  "correct": 557,
  "accuracy": 0.422289613343442,
  "finished": 1276,
  "finish_rate": 0.9673995451099318,
  "mean_completion_tokens": 191.71645185746777
}
saved item-level results -> outputs/evals/grid_math/glean_keep25_s1226_step100_chat.json
{
  "correct": 575,
  "accuracy": 0.4359363153904473,
  "finished": 1279,
  "finish_rate": 0.9696739954510993,
  "mean_completion_tokens": 193.40788476118271
}
saved item-level results -> outputs/evals/grid_math/glean_keep25_s1226_step150_chat.json