Changing one LLM agent's secret objective in Werewolf lowers its team's win rate and changes its reasoning, while its public chat stays deceptively normal.
Nobody likes a rat: On the willingness to report lies and the consequences thereof
1 Pith paper cite this work, alongside 57 external citations. Polarity classification is still indexing.
1
Pith paper citing it
57
external citations · OpenAlex
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Even More Deception: Objective Misalignment in Mixed-Motive LLM Multi-Agent Systems
Changing one LLM agent's secret objective in Werewolf lowers its team's win rate and changes its reasoning, while its public chat stays deceptively normal.