mzio/aprm-finqa_reasoning-gpt5m_med-gs4-s0-r2-train

--- {} --- # Act-PRM Rollout Dataset ## Run Metadata - **env_config**: `finqa/reasoning_gpt5m` - **model_config**: `oai_gpt5m_med` - **model**: `gpt-5-mini` - **split**: `train` - **group_size**: `4` - **seed**: `1` - **num_samples**: `62` - **num_trajectories**: `248` - **accuracy**: `41.5%` - **mean_reward**: `-0.169` - **run_cmd**: `main_api.py --env_config insurance/default_gpt5m --model_config oai_gpt5m_med --split train --group_size 4 --max_concurrent 5 --checkpoint_every 100 --dataset_name mzio/aprm-finqa_reasoning-gpt5m_med-gs8-s0-r2-train --seed 1 --verbose --env_config finqa/reasoning_gpt5m --group_size 4`