HF RL Explorer

OctoReasoner/bigcodebench_verl

BigCodeBench (verl eval set): RL dataset on Hugging Face. bigcode/bigcodebench v0.1.4 (1137 rows, data source=bigcodebench) plus bigcode/bigcodebench-hard v0.1.4 (146 rows, data source=bigcodebench hard; a subset of full, duplicated on purpose so full and hard pass@1 aggregate separately)…

OctoReasoner/bigcodebench_verl on the Hugging Face Hub