Prompt Injection Defense on Prompt Injection (Naive split)
0.05ASRStruQ
Evaluation Results
| Method | Links | |
|---|---|---|
| StruQFine-tuning attack source=Ignore attack2024.11 | 0.05 | |
| Ignore Defense2024.11 | 0.05 | |
| StruQFine-tuning attack source=Naive attack2024.11 | 0.5 | |
| None2024.11 | 10.55 |