List your company

Safety Evaluation on MultiJail Italian

100Safe Response RateDirect Instruction
7.85631.77855.779.622Sep 11, 2025

Evaluation results

MethodLinks
Direct InstructionModel=GPT-OSS-120b2025.09100
AIMModel=GPT-OSS-120b2025.09100
SteerMoEModel=GPT-OSS-120b2025.0990.2
SteerMoE + AIMModel=GPT-OSS-120b2025.0911.4