Gespeichert in:
| Hauptverfasser: | , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | https://arxiv.org/abs/2605.00460 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866911638386376704 |
|---|---|
| author | Jin, Weifei Wang, Xilong Zou, Wei Jia, Jinyuan Gong, Neil |
| author_facet | Jin, Weifei Wang, Xilong Zou, Wei Jia, Jinyuan Gong, Neil |
| contents | Retrieval-augmented generation (RAG) is vulnerable to prompt injection attacks, in which an adversary inserts malicious documents containing carefully crafted injected prompts into the knowledge database. When a user issues a question targeted by the attack, the RAG system may retrieve these malicious documents, whose injected prompts mislead it into generating attacker-specified answers, thereby compromising the integrity of the RAG system. In this work, we propose CleanBase, a method to detect malicious documents within a knowledge database. Our key insight is that malicious documents crafted for the same attack-targeted questions often exhibit high semantic similarity, as attackers deliberately make them consistent to improve attack success rates. Accordingly, CleanBase constructs a similarity graph over the knowledge database, where each node represents a document and an edge connects two nodes if their semantic similarity--computed using an embedding model--exceeds a statistically determined threshold. Due to their inherent similarity, malicious documents tend to form cliques within this graph. CleanBase detects such cliques and flags the corresponding documents as malicious. We theoretically derive upper bounds on CleanBase's false positive and false negative rates and empirically validate its effectiveness. Experimental results across multiple datasets and prompt injection attacks demonstrate that CleanBase accurately detects malicious documents and effectively safeguards RAG systems. Our source code is available at https://github.com/WeifeiJin/CleanBase. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_00460 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | CleanBase: Detecting Malicious Documents in RAG Knowledge Databases Jin, Weifei Wang, Xilong Zou, Wei Jia, Jinyuan Gong, Neil Cryptography and Security Machine Learning Retrieval-augmented generation (RAG) is vulnerable to prompt injection attacks, in which an adversary inserts malicious documents containing carefully crafted injected prompts into the knowledge database. When a user issues a question targeted by the attack, the RAG system may retrieve these malicious documents, whose injected prompts mislead it into generating attacker-specified answers, thereby compromising the integrity of the RAG system. In this work, we propose CleanBase, a method to detect malicious documents within a knowledge database. Our key insight is that malicious documents crafted for the same attack-targeted questions often exhibit high semantic similarity, as attackers deliberately make them consistent to improve attack success rates. Accordingly, CleanBase constructs a similarity graph over the knowledge database, where each node represents a document and an edge connects two nodes if their semantic similarity--computed using an embedding model--exceeds a statistically determined threshold. Due to their inherent similarity, malicious documents tend to form cliques within this graph. CleanBase detects such cliques and flags the corresponding documents as malicious. We theoretically derive upper bounds on CleanBase's false positive and false negative rates and empirically validate its effectiveness. Experimental results across multiple datasets and prompt injection attacks demonstrate that CleanBase accurately detects malicious documents and effectively safeguards RAG systems. Our source code is available at https://github.com/WeifeiJin/CleanBase. |
| title | CleanBase: Detecting Malicious Documents in RAG Knowledge Databases |
| topic | Cryptography and Security Machine Learning |
| url | https://arxiv.org/abs/2605.00460 |