Jailbreaking Attack (Salami 5-shot) on JailbreakBench
80ASRSmooth LLM
Evaluation Results
| Method | Links | |
|---|---|---|
| Smooth LLMTarget Model=GPT-4o2026.04 | 80 | |
| Self ReminderTarget Model=GPT-4o2026.04 | 83 | |
| Avg. ASRTarget Model=GPT-4o2026.04 | 83.5 | |
| Input FilterTarget Model=GPT-4o2026.04 | 84 | |
| w/o DefenseTarget Model=GPT-4o2026.04 | 87 |