# RAG-Safety-Bench: Reliable Evaluation of Retrieval-Augmented LLM Safety

arXiv: https://arxiv.org/abs/2609.11758 · HTML: https://mlsanity.com/paper/2609.11758/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Privacy and security of RAG, one of 15 directions in Retrieval-augmented generation.
- Builds on 15 of its 27 references that are in the map.
- First posted on arXiv 10 Sep 2026.

## Placed in

- Retrieval-augmented generation / [Privacy and security of RAG](https://mlsanity.com/rag/privacy-and-security-of-rag/)
- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 15 of 15)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Retrieval Augmentation Reduces Hallucination in Conversation](https://mlsanity.com/paper/2104.07567/) (2021)
- [WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs](https://mlsanity.com/paper/2406.18495/) (2024)
- [Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attack](https://mlsanity.com/paper/2404.01833/) (2024)
- [Safety-Tuned LLaMAs: Lessons From Improving the Safety of Large Language Models that Follow Instructions](https://mlsanity.com/paper/2309.07875/) (2023)
- [SORRY-Bench: Systematically Evaluating Large Language Model Safety Refusal](https://mlsanity.com/paper/2406.14598/) (2024)
- [The Good and The Bad: Exploring Privacy Issues in Retrieval-Augmented Generation (RAG)](https://mlsanity.com/paper/2402.16893/) (2024)
- [The Assistant Axis: Situating and Stabilizing the Default Persona of Language Models](https://mlsanity.com/paper/2601.10387/) (2026)
- [Follow My Instruction and Spill the Beans: Scalable Data Extraction from Retrieval-Augmented Generation Systems](https://mlsanity.com/paper/2402.17840/) (2024)
- ["Glue pizza and eat rocks" -- Exploiting Vulnerabilities in Retrieval-Augmented Generative Models](https://mlsanity.com/paper/2406.19417/) (2024)
- [Poison-RAG: Adversarial Data Poisoning Attacks on Retrieval-Augmented Generation in Recommender Systems](https://mlsanity.com/paper/2501.11759/) (2025)
- [MIRAGE: A Metric-Intensive Benchmark for Retrieval-Augmented Generation Evaluation](https://mlsanity.com/paper/2504.17137/) (2025)
- [SafeSearch: Do Not Trade Safety for Utility in LLM Search Agents](https://mlsanity.com/paper/2510.17017/) (2025)
- [Engineering the RAG Stack: A Comprehensive Review of the Architecture and Trust Frameworks for Retrieval-Augmented Generation Systems](https://mlsanity.com/paper/2601.05264/) (2025)

## Abstract (arXiv)

Allowing large language models (LLMs) to retrieve information from a set of trusted documents can increase reliability and reduce hallucination. However, recent work has demonstrated that retrieval-augmented generation (RAG) can have unintended side effects on the overall safety of the generated responses, when prompted for harmful or dangerous content. A clearer understanding of the mechanisms leading to this result is needed, as increasing numbers of end users turn to RAG to incorporate corporate documents and knowledge bases into LLM-based systems. We introduce RAG-Safety-Bench, a benchmark to measure the safety impact of RAG on LLM models. By removing the confounding effect of retriever quality, and cleanly separating the problem into four conditions -- non-RAG, RAG with an oracle document containing the answer to the harmful request, RAG with documents related to the harmful request but without the specific answer, and RAG with random, safe documents -- the benchmark isolates the impacts of different factors in the observed safety degradation. We report results across five open-source LLMs, showing an inverse relationship between benign and unsafe capability, strong evidence that baseline safety guardrails do not lead to downstream safety guarantees in the RAG case, and model-specific support for previous findings that even benign documents can lead to unsafe generation in retrieval-enabled systems.