List your company

Multi-agent interaction and social reasoning on Werewolf MultiAgentBench

55.75Task PerformanceETI
24.040432.272740.50548.7373Apr 21, 2026

Evaluation results

MethodLinks
ETIAgent=QWEN, Trait Source=QWEN2026.0455.7565.2
ETIAgent=QWEN, Trait Source=GPT2026.0449.9759.52
QWENTrait Source=none2026.0443.2860.2
ETIAgent=GPT, Trait Source=QWEN2026.0436.4657.56
ETIAgent=GPT, Trait Source=GPT2026.0429.5455.56
GPTTrait Source=none2026.0425.2654.32