Jailbreak Attack on VBCDE
2.7ASRRing-A-Bell
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Ring-A-BellAttack Protocol=Re-use attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 2.7 | 0.209 | 0.509 | — | |
| Ring-A-BellAttack Protocol=One-time attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 3.3 | 0.208 | 0.507 | — | |
| Chain-of-AttackAttack Protocol=Re-use attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 4 | 0.207 | 0.532 | — | |
| Chain-of-AttackAttack Protocol=One-time attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 4.7 | 0.211 | 0.52 | 17.72 | |
| MMADiffusionAttack Protocol=Re-use attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 8 | 0.223 | 0.549 | — | |
| DACAAttack Protocol=Re-use attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 8.3 | 0.226 | 0.551 | — | |
| MMADiffusionAttack Protocol=One-time attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 9.3 | 0.227 | 0.553 | — | |
| DACAAttack Protocol=One-time attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 9.7 | 0.228 | 0.56 | — | |
| SneakyPromptAttack Protocol=Re-use attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 9.7 | 0.235 | 0.596 | — | |
| SneakyPromptAttack Protocol=One-time attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 12.3 | 0.242 | 0.589 | 34.35 | |
| InceptionAttack Protocol=Re-use attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 26.3 | 0.243 | 0.614 | — | |
| InceptionAttack Protocol=One-time attack, Safety Filter=OpenAI input/output moderators, Memory Manager=BufferMem2025.04 | 32.3 | 0.247 | 0.636 | 12.18 |