List your company

Safety Evaluation on SafetyLLM

95.4Task PerformanceCausal Detoxify
88.234490.094791.95593.8153Dec 22, 2025

Evaluation results

MethodLinks
Causal DetoxifyAttack Method=Causal Detoxify2025.1295.40.82436.760.0277
AdaptiveAttack Method=Adaptive Poison2025.1294.250.925717.560.1666
CleanAttack Method=Clean2025.1293.1—4.73—
Two-StepAttack Method=Two-Step2025.1293.10.824314.870.1213
Extreme PoisonAttack Method=Extreme Poison2025.1292.340.959722.970.1945
OverpoisonAttack Method=Overpoison2025.1288.890.952724.320.1944
FusionAttack Method=Fusion2025.1288.510.547319.590.2218