List your company

Roleplay Personalization on Roleplay Human Evaluation (50 users, 11 questions)

72.3WinrateFSPO
68.03669.14370.2571.357Feb 26, 2025

Evaluation results

MethodLinks
FSPOBaseline Method=SFT2025.0272.3
FSPOBaseline Method=Base2025.0268.2