GuardAgent: Safeguard LLM Agents by a Guard Agent via Knowledge-Enabled Reasoning

Xiang, Zhen, Zheng, Linzhi, Li, Yanjie · arXiv · 2024