{"data":{"kind":"file","path":"README.md","version_id":"q49ht1uzjzd1lwog651zy8zf","entry":{"name":"README.md","path":"README.md","is_directory":false,"size":702,"modified_at":"2026-09-13T04:06:06.950000","content_hash":"19441127b94b343822230653f0e7559389ae82eec24aba95fbea8be23cecbaa6"},"entries":[],"content":"# i3-science\n\nScience tasks solved by an agent in a sandbox. Each task poses a science question with a boxed final answer; scored with math-verify against the gold answer, with an LLM-judge fallback when math-verify can't confirm it.\n\n## Taskset\n\n- **Source:** [PrimeIntellect/INTELLECT-3-RL](https://huggingface.co/datasets/PrimeIntellect/INTELLECT-3-RL)\n- **Size:** 29,307 tasks\n\n## Changelog\n\n- 2026-09-03: Restore default solver network access by reverting the `network_allow=[]` default-deny policy introduced in #780; training rollouts need outbound network.\n- 2026-08-31: Yield task records on demand so bounded evaluations construct only the requested prefix.\n- 2026-06-25: Initial v1 taskset.\n","encoding":"utf-8","truncated":false,"total_bytes":702},"status":null}