Spaces:

---

Foaster

/

Werewolf_benchmark

like 6

Running

Interesting benchmark, was eagerly awaiting for someone to use Mafia and Werewolf party games as a test for the social intelligence of LLMs. Great work!

Though if they are truly intelligent, they should be able to play the game at about the same reasoning skills in any natural language. Perhaps even in a constructed one.

Do you have a list of terms that should be translated and validated in order to play the games in languages other than English?