Furkejuvvon:
| Váldodahkkit: | , , , |
|---|---|
| Materiálatiipa: | Preprint |
| Almmustuhtton: |
2024
|
| Fáttát: | |
| Liŋkkat: | https://arxiv.org/abs/2402.06363 |
| Fáddágilkorat: |
Lasit fáddágilkoriid
Eai fáddágilkorat, Lasit vuosttaš fáddágilkora!
|
| _version_ | 1866929515226202112 |
|---|---|
| author | Chen, Sizhe Piet, Julien Sitawarin, Chawin Wagner, David |
| author_facet | Chen, Sizhe Piet, Julien Sitawarin, Chawin Wagner, David |
| contents | Recent advances in Large Language Models (LLMs) enable exciting LLM-integrated applications, which perform text-based tasks by utilizing their advanced language understanding capabilities. However, as LLMs have improved, so have the attacks against them. Prompt injection attacks are an important threat: they trick the model into deviating from the original application's instructions and instead follow user directives. These attacks rely on the LLM's ability to follow instructions and inability to separate prompts and user data. We introduce structured queries, a general approach to tackle this problem. Structured queries separate prompts and data into two channels. We implement a system that supports structured queries. This system is made of (1) a secure front-end that formats a prompt and user data into a special format, and (2) a specially trained LLM that can produce high-quality outputs from these inputs. The LLM is trained using a novel fine-tuning strategy: we convert a base (non-instruction-tuned) LLM to a structured instruction-tuned model that will only follow instructions in the prompt portion of a query. To do so, we augment standard instruction tuning datasets with examples that also include instructions in the data portion of the query, and fine-tune the model to ignore these. Our system significantly improves resistance to prompt injection attacks, with little or no impact on utility. Our code is released at https://github.com/Sizhe-Chen/StruQ. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2402_06363 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | StruQ: Defending Against Prompt Injection with Structured Queries Chen, Sizhe Piet, Julien Sitawarin, Chawin Wagner, David Cryptography and Security Recent advances in Large Language Models (LLMs) enable exciting LLM-integrated applications, which perform text-based tasks by utilizing their advanced language understanding capabilities. However, as LLMs have improved, so have the attacks against them. Prompt injection attacks are an important threat: they trick the model into deviating from the original application's instructions and instead follow user directives. These attacks rely on the LLM's ability to follow instructions and inability to separate prompts and user data. We introduce structured queries, a general approach to tackle this problem. Structured queries separate prompts and data into two channels. We implement a system that supports structured queries. This system is made of (1) a secure front-end that formats a prompt and user data into a special format, and (2) a specially trained LLM that can produce high-quality outputs from these inputs. The LLM is trained using a novel fine-tuning strategy: we convert a base (non-instruction-tuned) LLM to a structured instruction-tuned model that will only follow instructions in the prompt portion of a query. To do so, we augment standard instruction tuning datasets with examples that also include instructions in the data portion of the query, and fine-tune the model to ignore these. Our system significantly improves resistance to prompt injection attacks, with little or no impact on utility. Our code is released at https://github.com/Sizhe-Chen/StruQ. |
| title | StruQ: Defending Against Prompt Injection with Structured Queries |
| topic | Cryptography and Security |
| url | https://arxiv.org/abs/2402.06363 |