Jailbreak on AdvBench target: o4-mini (ASR)
64ASR (o4-mini)AGR
Evaluation Results
| Method | Links | |
|---|---|---|
| AGROptimization source=Qwen3-8B2026.05 | 64 | |
| AutoRANOptimization source=Qwen3-8B2026.05 | 50 | |
| H-CoTOptimization source=Qwen3-8B2026.05 | 43.3 | |
| ReNeLLMOptimization source=Qwen3-8B2026.05 | 23.3 | |
| PAIROptimization source=Qwen3-8B2026.05 | 18.7 | |
| AutoDANOptimization source=Qwen3-8B2026.05 | 16 | |
| GCGOptimization source=Qwen3-8B2026.05 | 2 |