Metraipolis › archive › dc_e2fb66947d34
RAG-PIBench: A Leakage-Aware Benchmark for Prompt-Injection Detection in Trustworthy RAG Systems — arXiv:2610.08571
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.08571
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T11:21:58.285Z
- SHA-256 of the archived text
ca9c53de31cde5942f6930332997602cf4d278187bddbe6326ccddbd34f17a73- Stable URL of this page
https://metraipolis.com/documento/dc_e2fb66947d34
How to cite
RAG-PIBench: A Leakage-Aware Benchmark for Prompt-Injection Detection in Trustworthy RAG Systems — arXiv:2610.08571 — arXiv (en), https://arxiv.org/abs/2610.08571 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_e2fb66947d34 (Metraipolis archive dc_e2fb66947d34, sha256 ca9c53de31cde5942f6930332997602cf4d278187bddbe6326ccddbd34f17a73)
BibTeX
@misc{dce2fb66947d34,
title = {RAG-PIBench: A Leakage-Aware Benchmark for Prompt-Injection Detection in Trustworthy RAG Systems — arXiv:2610.08571},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.08571},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_e2fb66947d34; sha256 ca9c53de31cde5942f6930332997602cf4d278187bddbe6326ccddbd34f17a73},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.08571. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Retrieval-Augmented Generation (RAG) systems are vulnerable to prompt-injection attacks embedded in retrieved content. We introduce RAG-PIBench, a benchmark for RAG-style prompt-injection detection containing 4,876 contextual examples across frozen train, validation, and protected-test splits. Using a leakage-aware construction pipeline and strict evaluation protocol, we compare keyword-based, semantic-reference, TF-IDF, and transformer-based detectors. DistilBERT achieves the best protected-test performance (F1 = 0.896, PR-AUC = 0.968), while TF-IDF SVM and logistic regression remain competitive. Our results demonstrate the value of leakage-aware benchmark design and strong sparse baselines for reliable prompt-injection detection in RAG systems. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Cryptography and Security (cs.CR) ; Artificial Intelligence (cs.AI); Machine Learning (cs.LG). Pagina del lavoro: https://arxiv.org/abs/2610.08571 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T11:15:24.845Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).