Jailbreak Attack on AdvBench (StrongREJECT score)
98StrongREJECT ScoreDictAttack
Evaluation Results
| Method | Links | |
|---|---|---|
| DictAttackModel=gemini-2.5-pro2025.03 | 98 | |
| AutoDAN-TurboModel=gemini-2.5-pro2025.03 | 87 | |
| TAPModel=gemini-2.5-pro2025.03 | 23 | |
| PAIRModel=gemini-2.5-pro2025.03 | 10 |