activity
20242026
most citedReliable Poisoned Sample Detection against Backdoor Attacks Enhanced by Sharpness Aware Minimization

1 citations · 1 across the 8 of their papers we have counts for

collaborators

8 papers

cs.CV2026

Distributed Implicit Harm: A Compositional Safety Blind Spot in MLLM-Based Video Moderation

Ruotong Wang, Zihao Zhu, Siwei Lyu +2

Despite their growing use in video moderation, multimodal large language models (MLLMs) exhibit a compositional safety blind spot: videos composed of seemingly benign components ca…

cs.CL2026

Can Induced Emotion Bias LLM Behaviors in Sequential Decision Making?

Minh Khoi Ho, Zihao Zhu, Runchuan Zhu +4

As Large Language Models (LLMs) are increasingly deployed as autonomous agents in high-stakes domains, understanding contextual factors that may modulate their decision-making beco…

cs.CV2026

BrandFusion: A Multi-Agent Framework for Seamless Brand Integration in Text-to-Video Generation

Zihao Zhu, Ruotong Wang, Siwei Lyu +2

The rapid advancement of text-to-video (T2V) models has revolutionized content creation, yet their commercial potential remains largely untapped. We introduce, for the first time,…

cs.LG2026

Unveiling Covert Toxicity in Multimodal Data via Toxicity Association Graphs: A Graph-Based Metric and Interpretable Detection Framework

Guanzong Wu, Zihao Zhu, Siwei Lyu +1

Detecting toxicity in multimodal data remains a significant challenge, as harmful meanings often lurk beneath seemingly benign individual modalities: only emerging when modalities…

cs.AI2025

AdvChain: Adversarial Chain-of-Thought Tuning for Robust Safety Alignment of Large Reasoning Models

Zihao Zhu, Xinyu Wu, Gehan Hu +3

Large Reasoning Models (LRMs) have demonstrated remarkable capabilities in complex problem-solving through Chain-of-Thought (CoT) reasoning. However, the multi-step nature of CoT i…

cs.CL2025

To Think or Not to Think: Exploring the Unthinking Vulnerability in Large Reasoning Models

Zihao Zhu, Hongbao Zhang, Ruotong Wang +3

Large Reasoning Models (LRMs) are designed to solve complex tasks by generating explicit reasoning traces before producing final answers. However, we reveal a critical vulnerabilit…