Jailbreak Attack on MI (MaliciousInstructions)
1QSRPuzzler
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PuzzlerTested Model=GPT3.52024.02 | 1 | 0.9 | |
| PuzzlerTested Model=GPT42024.02 | 1 | 0.86 | |
| PuzzlerTested Model=GPT4-Turbo2024.02 | 1 | 0.87 | |
| CRTested Model=GPT3.52024.02 | 0.9 | 0.936 | |
| PuzzlerTested Model=Gemini-pro2024.02 | 0.83 | 0.867 | |
| TCTested Model=GPT3.52024.02 | 0.53 | 0.867 | |
| CRTested Model=Gemini-pro2024.02 | 0.47 | 0.893 | |
| SIMUTested Model=GPT3.52024.02 | 0.4 | 0.909 | |
| TCTested Model=GPT42024.02 | 0.4 | 0.85 | |
| TAPTested Model=GPT3.52024.02 | 0.37 | 0.81 | |
| PuzzlerTested Model=LLama2-13B-chat2024.02 | 0.29 | 1 | |
| TAPTested Model=GPT42024.02 | 0.26 | 0.769 | |
| SIMUTested Model=Gemini-pro2024.02 | 0.17 | 0.882 | |
| TAPTested Model=Gemini-pro2024.02 | 0.14 | 0.786 | |
| CRTested Model=GPT42024.02 | 0.13 | 0.846 | |
| TAPTested Model=GPT4-Turbo2024.02 | 0.13 | 0.846 | |
| TCTested Model=GPT4-Turbo2024.02 | 0.07 | 0.857 | |
| PuzzlerTested Model=LLama2-7B-chat2024.02 | 0.03 | 0.667 | |
| TAPTested Model=LLama2-13B-chat2024.02 | 0.02 | 1 | |
| SIMUTested Model=GPT42024.02 | 0 | 0 | |
| CRTested Model=GPT4-Turbo2024.02 | 0 | 0 | |
| SIMUTested Model=GPT4-Turbo2024.02 | 0 | 0 | |
| TCTested Model=Gemini-pro2024.02 | 0 | 0 | |
| TAPTested Model=LLama2-7B-chat2024.02 | 0 | 0 | |
| CRTested Model=LLama2-7B-chat2024.02 | 0 | 0 | |
| TCTested Model=LLama2-7B-chat2024.02 | 0 | 0 | |
| SIMUTested Model=LLama2-7B-chat2024.02 | 0 | 0 | |
| CRTested Model=LLama2-13B-chat2024.02 | 0 | 0 | |
| TCTested Model=LLama2-13B-chat2024.02 | 0 | 0 | |
| SIMUTested Model=LLama2-13B-chat2024.02 | 0 | 0 |