Kunlun Zhu

Examining vulnerabilities and safety strategies for LLM-powered scientific agents.

A new framework enhances evaluation of RAG systems in specialized domains.

Large language models enhance collaboration in scientific research.