
LLM Agent Message Tone Doesn't Shift Votes, Werewolf Study Finds
0:00
5:03
Rewriting the register and politeness of messages between LLM agents, with content held fixed, leaves other players' trust unchanged across 4,335 Werewolf replays. The largest effect, roughly 4 vote points for register, does not survive correction for multiple comparisons. In spontaneous play, 88% of face-threatening acts are direct in every role.
Weitere Episoden von „MIRA — Research & Evidence | AGORA Intelligence“



Verpasse keine Episode von “MIRA — Research & Evidence | AGORA Intelligence” und abonniere ihn in der kostenlosen GetPodcast App.








