List your company

Safety Evaluation on WildGuard

25.5WildGuard Refusal RateBase
22.5242.63562.7582.865May 7, 2026

Evaluation results

MethodLinks
BaseModel=Mistral-7B2026.0525.559.5
BaseModel=Qwen3-0.6B2026.053252
TrigModel=Qwen3-0.6B, Sparsity=81.05%2026.0539.541.5
BaseModel=DeepSeek-8B2026.0547.538.5
TrigModel=DeepSeek-8B, Sparsity=81.38%2026.0556.530.5
TrigModel=Mistral-7B, Sparsity=59.34%2026.056921.5
TrigModel=Vicuna-7B, Sparsity=84.40%2026.0574.523
BaseModel=Vicuna-7B2026.0576.520.5
LCDDModel=Qwen3-0.6B, Sparsity=81.05%2026.0588.53.5
SFTModel=Qwen3-0.6B2026.05942
LCDDModel=DeepSeek-8B, Sparsity=81.38%2026.0596.52
LCDDModel=Mistral-7B, Sparsity=59.34%2026.0597.51
LCDDModel=Vicuna-7B, Sparsity=84.40%2026.05981.5
SFTModel=Vicuna-7B2026.05990
SFTModel=DeepSeek-8B2026.051000
SFTModel=Mistral-7B2026.051000