Dataset and Lessons Learned from the 2024 SaTML LLM Capture-the-Flag Competition
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2024
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866916283579105280 |
|---|---|
| author | Debenedetti, Edoardo Rando, Javier Paleka, Daniel Florin, Silaghi Fineas Albastroiu, Dragos Cohen, Niv Lemberg, Yuval Ghosh, Reshmi Wen, Rui Salem, Ahmed Cherubin, Giovanni Zanella-Beguelin, Santiago Schmid, Robin Klemm, Victor Miki, Takahiro Li, Chenhao Kraft, Stefan Fritz, Mario Tramèr, Florian Abdelnabi, Sahar Schönherr, Lea |
| author_facet | Debenedetti, Edoardo Rando, Javier Paleka, Daniel Florin, Silaghi Fineas Albastroiu, Dragos Cohen, Niv Lemberg, Yuval Ghosh, Reshmi Wen, Rui Salem, Ahmed Cherubin, Giovanni Zanella-Beguelin, Santiago Schmid, Robin Klemm, Victor Miki, Takahiro Li, Chenhao Kraft, Stefan Fritz, Mario Tramèr, Florian Abdelnabi, Sahar Schönherr, Lea |
| contents | Large language model systems face important security risks from maliciously crafted messages that aim to overwrite the system's original instructions or leak private data. To study this problem, we organized a capture-the-flag competition at IEEE SaTML 2024, where the flag is a secret string in the LLM system prompt. The competition was organized in two phases. In the first phase, teams developed defenses to prevent the model from leaking the secret. During the second phase, teams were challenged to extract the secrets hidden for defenses proposed by the other teams. This report summarizes the main insights from the competition. Notably, we found that all defenses were bypassed at least once, highlighting the difficulty of designing a successful defense and the necessity for additional research to protect LLM systems. To foster future research in this direction, we compiled a dataset with over 137k multi-turn attack chats and open-sourced the platform. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_07954 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Dataset and Lessons Learned from the 2024 SaTML LLM Capture-the-Flag Competition Debenedetti, Edoardo Rando, Javier Paleka, Daniel Florin, Silaghi Fineas Albastroiu, Dragos Cohen, Niv Lemberg, Yuval Ghosh, Reshmi Wen, Rui Salem, Ahmed Cherubin, Giovanni Zanella-Beguelin, Santiago Schmid, Robin Klemm, Victor Miki, Takahiro Li, Chenhao Kraft, Stefan Fritz, Mario Tramèr, Florian Abdelnabi, Sahar Schönherr, Lea Cryptography and Security Artificial Intelligence Large language model systems face important security risks from maliciously crafted messages that aim to overwrite the system's original instructions or leak private data. To study this problem, we organized a capture-the-flag competition at IEEE SaTML 2024, where the flag is a secret string in the LLM system prompt. The competition was organized in two phases. In the first phase, teams developed defenses to prevent the model from leaking the secret. During the second phase, teams were challenged to extract the secrets hidden for defenses proposed by the other teams. This report summarizes the main insights from the competition. Notably, we found that all defenses were bypassed at least once, highlighting the difficulty of designing a successful defense and the necessity for additional research to protect LLM systems. To foster future research in this direction, we compiled a dataset with over 137k multi-turn attack chats and open-sourced the platform. |
| title | Dataset and Lessons Learned from the 2024 SaTML LLM Capture-the-Flag Competition |
| topic | Cryptography and Security Artificial Intelligence |
| url | https://arxiv.org/abs/2406.07954 |