作者位次 #2/5ACL 2026Findings· Findings of ACL
Measuring Large Language Models’ Adversarial Behavior in Social Deduction Games
Marissa Zhao Li · Esha Shivakumar · Peiran Wang · Ying Li · Yuan Tian
摘要
As large language models are increasingly adopted and trusted in real-world applications, understanding their behavior beyond single-turn prompting has become critical. Existing safety evaluations primarily focus on refusal-based methods that test whether models avoid responding to inappropriate...