Xiaogeng Liu

Doxing via the Lens: Revealing Location-related Privacy Leakage on Multi-modal Large Reasoning Models

Computer-Use Agent Frameworks Can Expose Realistic Risks Through Tactics, Techniques, and Procedures

VLA-Risk: Benchmarking Vision-Language-Action Models with Physical Robustness

AutoDAN-Turbo: A Lifelong Agent for Strategy Self-Exploration to Jailbreak LLMs

DRIFT: Dynamic Rule-Based Defense with Injection Isolation for Securing LLM Agents

Can Watermarks be Used to Detect LLM IP Infringement For Free?

Prompt Injection Benchmark for Foundation Model Integrated Systems

MuirBench: A Comprehensive Benchmark for Robust Multi-image Understanding

MetaAgent: Automatically Constructing Multi-Agent Systems Based on Finite State Machines

AutoHijacker: Automatic Indirect Prompt Injection Against Black-box LLM Agents

MetaAgent: Automatically Building Multi-Agent System based on Finite State Machine