Xinfeng Li

AudioTrust: Benchmarking The Multifaceted Trustworthiness of Audio Large Language Models

Tug-of-War No More: Harmonizing Accuracy and Robustness in Vision-Language Models via Stability-Aware Task Vector Merging

OrthAlign: Orthogonal Subspace Decomposition for Non-Interfering Multi-Objective Alignment

Learning to Rank Chain-of-Thought: Using a Small Model

HAI-Eval: Measuring Human-AI Synergy in Collaborative Coding

A-MemGuard: A Proactive Defense Framework for LLM-Based Agent Memory

MedSentry: Understanding and Mitigating Safety Risks in Medical LLM Multi-Agent Systems

Can LLMs Refuse Questions They Do Not Know? Measuring Knowledge-Aware Refusal in Factual Tasks

GTD: Dynamic Generation of Multi LLM Agents Communication Topologies with Graph Diffusion Models

Energy-Driven Steering: Reducing False Refusals in Large Language Models

UniErase: Towards Balanced and Precise Unlearning in Language Models

GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoning

Adversarial Attacks against Closed-Source MLLMs via Feature Optimal Alignment

AgentAuditor: Human-level Safety and Security Evaluation for LLM Agents

SeCon-RAG: A Two-Stage Semantic Filtering and Conflict-Free Framework for Trustworthy RAG

Prompt-Independent Safe Decoding to Restrain Unsafe Image Generation for Text-to-Image Models against White-Box Adversary