DeepFact: Co-Evolving Benchmarks and Agents for Deep Research Factuality

1. 1Develop static expert-labeled benchmarks for fact-checking
2. 2Evaluate the brittleness of these benchmarks in the context of deep research reports
3. 3Design co-evolving benchmarks and agents to improve factuality verification
4. 4Test the transferability of fact-checkers to deep research reports

AI researchers and developers working on LLM agents and fact-checking systems can benefit from DeepFact, as it addresses the challenge of verifying claim-level factuality in deep research reports

💡 Static expert-labeled benchmarks are brittle for verifying factuality in deep research reports, requiring co-evolving benchmarks and agents

🔍 DeepFact: co-evolving benchmarks & agents for factuality in deep research reports