Distractor Selection on Tool selection benchmark
30Certified Upper Bound (pu)SecAlign
Evaluation Results
| Method | Links | |
|---|---|---|
| SecAlignMitigation=SecAlign, Selector model=Llama-3.1, Generator model=Gemma-32025.10 | 30 | |
| Anomaly mon.Mitigation=Anomaly mon., Selector model=Llama-3.1, Generator model=Gemma-32025.10 | 26 | |
| StruQMitigation=StruQ, Selector model=Llama-3.1, Generator model=Gemma-32025.10 | 24 | |
| PerplexityMitigation=Perplexity, Selector model=Llama-3.1, Generator model=Gemma-32025.10 | 23 | |
| Defended retr.Mitigation=Defended retr., Selector model=Llama-3.1, Generator model=Gemma-32025.10 | 22 | |
| NoneMitigation=None, Selector model=Llama-3.1, Generator model=Gemma-32025.10 | 20 |