List your company

Safety Evaluation on BeaverTails Unsafe Animal Abuse

85Safe RateSafeMoE
24.36840.10955.8571.591May 30, 2026

Evaluation results

MethodLinks
SafeMoEBase Model=Mistral-7B, Variant=MoE-XL2026.05856.24
SafeMoEBase Model=Qwen-3B, Variant=MoE-XL2026.05606.71
Mistral-7BTuning=Standard instruction-tuned2026.0533.33.44
Qwen-3BTuning=Standard instruction-tuned2026.0526.76