1 paper
Jaewoo Ahn, Junseo Kim, Hyunseo Kim +4
Strategic deception by LLM and VLM agents has emerged as a central AI alignment and safety concern. Social-deduction games (where each player holds a hidden role and communicates w…