Gespeichert in:
| Hauptverfasser: | , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2601.09292 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866915728421027840 |
|---|---|
| author | Dolcetti, Greta Zizzo, Giulio Maffeis, Sergio |
| author_facet | Dolcetti, Greta Zizzo, Giulio Maffeis, Sergio |
| contents | We present an experimental evaluation that assesses the robustness of four open source LLMs claiming function-calling capabilities against three different attacks, and we measure the effectiveness of eight different defences. Our results show how these models are not safe by default, and how the defences are not yet employable in real-world scenarios. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_09292 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Blue Teaming Function-Calling Agents Dolcetti, Greta Zizzo, Giulio Maffeis, Sergio Cryptography and Security Artificial Intelligence We present an experimental evaluation that assesses the robustness of four open source LLMs claiming function-calling capabilities against three different attacks, and we measure the effectiveness of eight different defences. Our results show how these models are not safe by default, and how the defences are not yet employable in real-world scenarios. |
| title | Blue Teaming Function-Calling Agents |
| topic | Cryptography and Security Artificial Intelligence |
| url | https://arxiv.org/abs/2601.09292 |