SafeRAG: Benchmarking Security in Retrieval-Augmented Generation of Large Language Model
SafeRAG: Бенчмаркинг безопасности в системах генерации с дополненной поиском для больших языковых моделей
2025-01-01
SCID: 54.1/p4nq6xhu
Discuss with AI
SafeRAGlarge language modelsretrieval-augmented generationsecurity benchmarking
Figures from the paper
Abstract (AI)
Xun Liang, Simin Niu, Zhiyu Li, Sensen Zhang, Hanyu Wang, Feiyu Xiong, Zhaoxin Fan, Bo Tang, Jihao Zhao, Jiawei Yang, Shichao Song, Mengwei Wang. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.
Key Findings
1
SafeRAG focuses on systematically assessing vulnerabilities arising from retrieved information and its interaction with language-model generation.
2
The benchmark is presented as a basis for comparing RAG security across systems, though the provided abstract contains no quantitative results or specific attack categories.
3
The paper introduces SafeRAG, a benchmark for evaluating security risks in retrieval-augmented generation systems for large language models.
Research Object
retrieval-augmented generation systems using large language models
Research Subject
security vulnerabilities and robustness of retrieval-augmented generation systems under adversarial attacks
Publication Details
Publication Date
2025-01-01
Journal
Publisher
ISSN
Cited by
74
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest