List your company

Safety Evaluation on MultiJail English

100Safe Response RateAIM
5.8830.31554.7579.185Sep 11, 2025

Evaluation results

MethodLinks
AIMModel=GPT-OSS-120b2025.09100
Direct InstructionModel=GPT-OSS-120b2025.0999.7
SteerMoEModel=GPT-OSS-120b2025.0994.3
SteerMoE + AIMModel=GPT-OSS-120b2025.099.5