一作ACL 2026Findings· Findings of ACL
Towards Identification and Intervention of Safety-Critical Parameters in Large Language Models
Weiwei Qi · Zefeng Wu · Tianhang Zheng · Zikang Zhang · Xiaojun Jia · Zhan Qin · Kui Ren
摘要
Ensuring Large Language Model (LLM) safety is crucial, yet the lack of a clear understanding about safety mechanisms hinders the development of precise and reliable methodologies for safety intervention across diverse tasks. To better understand and control LLM safety, we propose the Expected Saf...