mondaycake/lifesci-harbor-bench
LifeSci-Harbor-Bench: 503 Open Research Tasks for Agentic Life Science 500 ultra-long-horizon, open-ended research tasks evaluating coding agents on real, computationally verifiable (dry-lab) research in the life sciences, with an emphasis on aging / longevity / healthspan. Every task is self-contained, derived from real scientific literature, clinical trials, and public databases, and scored through a v4.0 11-layer verifier (6-layer F1 scoring + 5-layer F2 anti-hacking) +… See the full description on the dataset page: https://huggingface.co/datasets/mondaycake/lifesci-harbor-bench.
| Author / organization | mondaycake |
|---|---|
| Downloads | 168 (as reported by Hugging Face) |
| Likes | 0 |
| Created | 6 September 2026 |
| Last updated at source | 14 September 2026 |
Popularity, last 3 days
168 now (-2 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.
Connected in the knowledge graph
- World Feedback for Clinical Agents: Diagnosing RL in FHIR EnvironmentsPaper · cites
- Agentic BAIM-LLM Evaluation (ABLE): Benchmarking LLM Use of Protein Design ToolsPaper · cites
- Fara-1.5: Scalable Learning Environments for Computer Use AgentsPaper · cites
- Cheap Reward Hacking DetectionPaper · cites
- Do Agent Benchmarks Measure Capability? Protocol Validity in the Age of Agentic AIPaper · cites
Related papers
- Cheap Reward Hacking DetectionPaper
- Fara-1.5: Scalable Learning Environments for Computer Use AgentsPaper
- World Feedback for Clinical Agents: Diagnosing RL in FHIR EnvironmentsPaper
- Do Agent Benchmarks Measure Capability? Protocol Validity in the Age of Agentic AIPaper
- Agentic BAIM-LLM Evaluation (ABLE): Benchmarking LLM Use of Protein Design ToolsPaper
Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.