arXiv:2606. 00686v1 Announce Type: new Abstract: The prevailing paradigm in large language model (LLM) alignment operates via erasure, filtering unsafe data or training models to strictly refuse harmful prompts.
arXiv:2608. 04565v1 Announce Type: cross Abstract: LLM-based search agents are widely used for information-seeking tasks, but their reliance on external tool returns introduces a critical security risk: web content retrieved during execution is untrusted, exposing agents to prompt injection and goal hijacking.
arXiv:2603. 00801v2 Announce Type: replace Abstract: Language agents increasingly act as web-enabled systems that search, browse, and synthesize information from diverse sources.
RAG-Safety-Bench is a benchmark designed to evaluate how retrieval-augmented generation (RAG) affects the safety of large language models (LLMs). It isolates safety impacts by testing four conditions: non-RAG, RAG with an oracle document, RAG with related but non-answer documents, and RAG with random safe documents. Results on five open-source LLMs reveal an inverse relationship between benign and unsafe capabilities, show that baseline safety guardrails do not guarantee safety in RAG, and confirm that even benign documents can trigger unsafe generation.
arXiv:2610.23093v1 Announce Type: new Abstract: LLM agents increasingly use retrieval tools across languages, yet their safety evaluation remains English-centric. It remains unclear whether language...
RAG-Safety-Bench is a benchmark designed to evaluate how retrieval-augmented generation (RAG) affects the safety of large language models (LLMs). It isolates safety impacts by testing four conditions: non-RAG, RAG with an oracle document, RAG with related but non-answer documents, and RAG with random safe documents. Results on five open-source LLMs reveal an inverse relationship between benign and unsafe capabilities, show that standard safety guardrails do not guarantee safety in RAG, and confirm that even benign documents can trigger unsafe outputs.
The paper surveys attacks and defenses in Retrieval-Augmented Generation (RAG), a technique that improves large language models by grounding outputs in external knowledge. It identifies new robustness and security risks such as corpus poisoning, backdoor attacks, privacy leakage, and fairness violations, and notes that existing surveys inadequately cover attacker objectives, threat models, and stage-specific defenses. The survey offers a unified, pipeline-aware overview, formalizing threat models across the corpus, retriever, and generator, categorizing attacks by accuracy, privacy, and fairness, and reviewing defenses for retrieval, rerank, generation, and traceback stages, while also summarizing robustness benchmarks and explainability methods.
arXiv:2406.00083v3 Announce Type: replace-cross Abstract: Retrieval-Augmented Generation (RAG) enhances Large Language Models (LLMs) by retrieving relevant information from external knowledge bases t...
arXiv:2607. 18295v1 Announce Type: new Abstract: We study whether alignment schemes that reshape a base model's output distribution, combined with bounded safety filters, can drive the probability of harmful behavior to zero in modern large language models.
Large language models (LLMs) are rigorously aligned to refuse harmful requests, a process that inherently cultivates a latent capacity to evaluate and recognize unsafe content. In this work, we reveal that this advanced safety awareness inadvertently introduces a fatal vulnerability.
arXiv:2606. 18356v1 Announce Type: cross Abstract: Tool-using language-model agents introduce security failures that go beyond unsafe text: they can disclose protected objects, write persistent memory, send messages, modify databases, or trigger harmful code and tool effects.
DeepSafe‑SciHazard introduces a real‑world‑grounded benchmark for measuring scientific safety risks, comprising 2,400 hazardous and 600 oversafety questions across 12 disciplines. It defines a decomposed harm scoring system, DeHarm‑Score, that evaluates query severity, refusal behavior, and response‑level risk, further breaking down harm into executability and net‑new risk. An expert‑validation study shows DeHarm‑Score aligns 90.17% better with expert judgments than the best baseline, and the benchmark reveals that deep research agents score 32.3% higher on mean DeHarm‑Score than standard LLMs, highlighting a critical blind spot in current safety defenses.