Jailbreak Attack on AdvBench GPT-4 1.0 (test)
0.77ASRMAC
Evaluation Results
| Method | Links | |
|---|---|---|
| MACAttack Type=Automated, Evaluator=GPT-42025.12 | 0.77 | |
| COLD-AttackAttack Type=Automated, Evaluator=GPT-42025.12 | 0.77 | |
| Base64Attack Type=Automated, Evaluator=GPT-42025.12 | 0.77 | |
| PromptAttackAttack Type=Manual, Evaluator=GPT-42025.12 | 0.96 | |
| GCGAttack Type=Automated, Evaluator=GPT-42025.12 | 1.73 | |
| ArtPromoptAttack Type=Manual, Evaluator=GPT-42025.12 | 1.75 | |
| CodeChameleonAttack Type=Manual, Evaluator=GPT-42025.12 | 22.27 | |
| AutoDANAttack Type=Automated, Evaluator=GPT-42025.12 | 26.54 | |
| PAIRAttack Type=Automated, Evaluator=GPT-42025.12 | 27.18 | |
| DeepInceptionAttack Type=Manual, Evaluator=GPT-42025.12 | 27.27 | |
| TAPAttack Type=Automated, Evaluator=GPT-42025.12 | 40.97 | |
| SelfCipherAttack Type=Manual, Evaluator=GPT-42025.12 | 41.73 | |
| GPTFuzzerAttack Type=Automated, Evaluator=GPT-42025.12 | 42.5 | |
| ReNeLLMAttack Type=Manual, Evaluator=GPT-42025.12 | 68.08 | |
| DRAAttack Type=Manual, Evaluator=GPT-42025.12 | 82.88 | |
| FlipattackAttack Type=Manual, Evaluator=GPT-42025.12 | 86.73 | |
| EquaCodeEvaluator=GPT-42025.12 | 91.92 |