1
AIR: Improving Agent Safety through Incident Response
不同于传统侧重预防的AI安全机制，这篇论文提出通过事故响应、遏制与恢复来提升LLM代理的安全性。
arXiv:2602.11749v2 Announce Type: replace Abstract: Large Language Model (LLM) agents are increasingly deployed in practice across a wide range of aut…