Jailbreak Defense Efficiency and Effectiveness on JailbreakBench Summary Aggregates
1Avg ASRTrapSuffix
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| TrapSuffix2026.02 | 1 | 0 | 15.87 | |
| WildGuardDeployment Stage=Post2026.02 | 1 | 2.23 | 29,013 | |
| SmoothLLM2026.02 | 23 | 38.83 | 2,185 | |
| PPL2026.02 | 36 | 0.09 | 13,120 | |
| GradSafe2026.02 | 51 | 2.37 | 8,744 |