Sponge Attack Detection on AUTO-DOS
100F1 ScoreSHIELD
Evaluation Results
| Method | Links | |
|---|---|---|
| SHIELDTarget Model=LLaMA22026.01 | 100 | |
| Harm-filterTarget Model=LLaMA22026.01 | 87.57 | |
| Perplexity-filterTarget Model=LLaMA22026.01 | 36.51 |
| Method | Links | |
|---|---|---|
| SHIELDTarget Model=LLaMA22026.01 | 100 | |
| Harm-filterTarget Model=LLaMA22026.01 | 87.57 | |
| Perplexity-filterTarget Model=LLaMA22026.01 | 36.51 |