7 papers · 1 filter
Your Agentic LLMs Secretly Encode Indirect Prompt-Injection Exposure in Hidden States
Jianshuo Dong, Yiming Liu, Maosen Zhang +6
Agentic LLMs are vulnerable to indirect prompt injection (IPI) attacks, e.g., malicious side-tasks hidden in external tool results. While many efforts have sought to address this t…
The Model's Tell: Measuring Context-Leakage Attack Signals with Behavior Gauges
Maosen Zhang, Jianshuo Dong, Boting Lu +5
LLMs increasingly rely on external contexts, such as pre-defined system prompts or retrieved documents, to improve generation quality. However, processing these contexts alongside…
LeakDojo: Decoding the Leakage Threats of RAG Systems
Maosen Zhang, Jianshuo Dong, Boting Lu +4
Retrieval-Augmented Generation (RAG) enables large language models (LLMs) to leverage external knowledge, but also exposes valuable RAG databases to leakage attacks. As RAG systems…
DREAM: Scalable Red Teaming for Text-to-Image Generative Systems via Distribution Modeling
Boheng Li, Junjie Wang, Yiming Li +7
Despite the integration of safety alignment and external filters, text-to-image (T2I) generative systems are still susceptible to producing harmful content, such as sexual or viole…
Can Large Language Models Automate the Refinement of Cellular Network Specifications?
Jianshuo Dong, Yuanjie Li, Jun Liu +2
Cellular networks, e.g., 4G/5G, rely on complex technical specifications to ensure correct functionality; however, these specifications often contain flaws or ambiguities. In this…
An Engorgio Prompt Makes Large Language Model Babble on
Jianshuo Dong, Ziyuan Zhang, Qingjie Zhang +7
Auto-regressive large language models (LLMs) have yielded impressive performance in many real-world tasks. However, the new paradigm of these LLMs also exposes novel threats. In th…
One-bit Flip is All You Need: When Bit-flip Attack Meets Model Training
Jianshuo Dong, Han Qiu, Yiming Li +5
Deep neural networks (DNNs) are widely deployed on real-world devices. Concerns regarding their security have gained great attention from researchers. Recently, a new weight modifi…