DeepEvidence: Knowledge-Graph-Driven Deep Research Benchmarks
Dataset Summary
DeepEvidence is a curated suite of evaluation-only benchmarks designed to assess AI agents and large language models on knowledge-graph-driven Deep Research for biomedicine.
Modern biomedical discovery relies on navigating vast, heterogeneous knowledge graphs spanning literature, genes, pathways, drugs, diseases, and clinical trials. While recent “Deep Research” agents demonstrate strong… See the full description on the dataset page: https://huggingface.co/datasets/zifeng-ai/DeepEvidence.
