SOURCE-LINKED INTELLIGENCE

Do LLMs Trust the Accuser or the Accusation? Measuring Belief Shifts in Werewolf

arXiv · AI, language, vision and robotics · article · Sep 11, 2026 · UTC

Social-deduction games such as Werewolf are increasingly used to evaluate LLM agents, but existing evaluations often rely on final game outcomes. We propose a belief-shift evaluation benchmark in Werewolf for analyzing communication skills through belief updating. Using LLM-played games, we annotate suspicion and accusation messages and measure how an observing village-side model's beliefs change after each message. We evaluate 40 open-weight LLM configurations on 1,224 annotated messages. Our results show that larger models better distinguish true wolves from villagers based on game history,

Read original source ↗ Open in workspace

recordType

paper

region

Global