MIRA — Research & Evidence | AGORA Intelligence podcast

LLM Agent Message Tone Doesn't Shift Votes, Werewolf Study Finds

0:00
5:03
15 Sekunden vorwärts
15 Sekunden vorwärts
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“