10 papers · 1 filter
SceneJail: Exploiting Video Scenario Context to Jailbreak Multimodal LLMs
Wenyu Chen, Li Wang, Chuanchao Zang +5
Video Multimodal Large Language Models (Video-MLLMs) support reasoning over video inputs, yet remain vulnerable to jailbreak attacks that elicit policy-violating responses. Existin…
Transferable End-to-End Optimization for Indirect Long-Term Memory Poisoning in LLM Agents
Chuanchao Zang, Jianing Wang, Wenyu Chen +5
Long-term memory can turn untrusted external content into persistent influence over an LLM agent's future decisions, creating the threat of indirect memory poisoning. A successful…
Extracting Knowledge from Tools in LLM Agents
Chuanchao Zang, Jianing Wang, Wenyu Chen +6
LLM agents commonly use knowledge-based tools and access their underlying files, databases, and search indexes through tool invocation. This integration improves agents' ability to…
Understanding Stage-Wise Utility-Risk Trade-offs in LLM Agent Memory
Chuanchao Zang, Zijian Cao, Xiangtao Meng +6
Long-term memory is becoming a core capability of LLM agents, enabling personalization and long-horizon interaction. However, memory mechanisms that retain, transform, or expose mo…
Isolated but Exposed: Persistence-Based Memory Extraction Attack on LLM Agents
Xinyu Gao, Wenyu Chen, Xiangtao Meng +5
LLM-based agents extend large language models with long-term memory (LTM) that persists privacy-sensitive user data across sessions. Production systems mitigate extraction risks th…
Defenses at Odds: Measuring and Explaining Defense Conflicts in Large Language Models
Xiangtao Meng, Wenyu Chen, Chuanchao Zang +5
Large language models (LLMs) may require additional defenses after deployment as risks and governance requirements evolve. Subsequent defenses can interact with earlier defenses, r…
Not All Tokens Are Created Equal: Query-Efficient Jailbreak Fuzzing for LLMs
Wenyu Chen, Xiangtao Meng, Chuanchao Zang +6
Large Language Models(LLMs) are widely deployed, yet are vulnerable to jailbreak prompts that elicit policy-violating outputs. Although prior studies have uncovered these risks, th…
VidLeaks: Membership Inference Attacks Against Text-to-Video Models
Li Wang, Wenyu Chen, Ning Yu +2
The proliferation of powerful Text-to-Video (T2V) models, trained on massive web-scale datasets, raises urgent concerns about copyright and privacy violations. Membership inference…
One Risk Down, Another Up: Cross-Risk Interactions Induced by LLM Defenses
Xiangtao Meng, Tianshuo Cong, Li Wang +4
Large Language Models (LLMs) are increasingly deployed in high-stakes settings, where they face diverse risks. Numerous defense strategies have been proposed to mitigate these risk…
Beyond Known Fakes: Generalized Detection of AI-Generated Images via Post-hoc Distribution Alignment
Li Wang, Wenyu Chen, Xiangtao Meng +2
The rapid proliferation of highly realistic AI-generated images poses serious security threats such as misinformation and identity fraud. Detecting generated images in open-world s…