ZY
Zonghao Ying
cs.CRcs.AIcs.CVcs.CLcs.ROcs.CYcs.LGcs.MA
On Valency
published · living versionsW_qn3dhfa8·v1 · currentpublished
VagueBench: Benchmarking the Robustness of Embodied Agents on Ambient Context Illusions
with Yang Qu, Qiang Fu, Jiayi Zhao, Yaoyuan Zhang +5
1 version
Preprints & journals
51 papers in the corpus · 2024–2026Hidden in Plain Sight: Advancing Harmful Content Understanding via Perception-Aware Adversarial Self-Play2610.15449v2 · Yisong Xiao, Benlei Cui, Haiwen Hong et al.2026 · 0 citationsarXiv
pikit: A Composable Toolkit for Indirect Prompt Injection Research and Evaluation2609.36817v1 · Zonghao Ying, Xiangfan Wu, Bo Yang et al.2026 · 0 citationsarXiv
CoDeL: Co-Evolutionary Defense against Indirect Prompt Injection in LLM-based Agents2609.34463v1 · Xiao Yang, Yangchen Ou, Yuhan Gao et al.2026 · 0 citationsarXiv
Collective Loss of Control in LLM Agent Systems: An Epidemic Account of Mutation, Contagion, and Recovery2609.18460v1 · Xiangfan Wu, Zonghao Ying, Huiyu Wu et al.2026 · 0 citationsarXiv
The Missing Boundary: How Autonomous Agents Lose Control2609.11024v2 · Zonghao Ying, Xiangfan Wu, Bo Yang et al.2026 · 0 citationsarXiv
Delegated Misalignment: How Multi-Agent Structures Amplify LLM Safety Risks2609.27900v1 · Zonghao Ying, Jiaqi Yan, Huize Luo et al.2026 · 0 citationsarXiv
Ventor-QTest: Threat-Model-Driven Verification of Vendor-Hosted LLM APIs2608.16391v1 · Xiangfan Wu, Zonghao Ying, Huiyu Wu et al.2026 · 0 citationsarXiv
Security Assessment of DeepSeek Harness with A.I.G: Evaluating Resistance to Indirect Prompt Injection2608.16393v2 · Zonghao Ying, Xiangfan Wu, Huiyu Wu et al.2026 · 0 citationsarXiv
Mitigating Bias in Large Vision-Language Models via Counterfactual Ensemble Decoding2608.21415v1 · Yisong Xiao, Aishan Liu, Yongxin Huang et al.2026 · 0 citationsarXiv
SkillSentry: Adaptive Honey Worlds for Dynamic Safety Testing of Agent Skills2608.03485v1 · Nizhang Li, Zonghao Ying, Xiangfan Wu et al.2026 · 0 citationsarXiv
SkillJack: Persistent Skill Backdoors in Self-Evolving Agents2608.03509v2 · Zonghao Ying, Xiangfan Wu, Huiyu Wu et al.2026 · 0 citationsarXiv
SafeFlow: Semantic Information-Flow Control for Blocking Malicious Propagation in Multi-Agent Systems2607.25255v2 · Haowen Dai, Zonghao Ying, Wenfeng Li et al.2026 · 0 citationsarXiv
SafeGen: Goal-Conditioned Video Diffusion of Safety-Critical Scenarios for VLM-Based Autonomous Driving2607.19701v1 · Jiangfan Liu, Zexuan Cui, Tianyuan Zhang et al.2026 · 0 citationsarXiv
Dynamic Defense Profiling Enables Cognitive Jailbreak of Text-to-Image Models2607.17779v1 · Dongdong Yang, Deyue Zhang, Zhao Liu et al.2026 · 0 citationsarXiv
Technical Report on the CVPR 2026@AdvML Workshop Challenge2607.11560v1 · Tianyuan Zhang, Zonglei Jing, Jiangfan Liu et al.2026 · 0 citationsarXiv
Securing the AI Agent: A Unified Framework for Multi-Layer Agent Red Teaming2606.31227v1 · Yong Yang, Xing Zheng, Huiyu Wu et al.2026 · 0 citationsarXiv
DMN: A Compositional Framework for Jailbreaking Multimodal LLMs with Multi-Image Inputs2605.18915v1 · Wenzhuo Xu, Zhipeng Wei, Zonghao Ying et al.2026 · 0 citationsarXiv
Attention Hijacking: Response Manipulation Across Queries in Vision-Language Models2605.17310v1 · Zhiqiang Wang, Dongrui Liu, Yan Li et al.2026 · 0 citationsarXiv
GuardAD: Safeguarding Autonomous Driving MLLMs via Markovian Safety Logic2605.10386v1 · Tianyuan Zhang, Peng Yue, Zihao Peng et al.2026 · 0 citationsarXiv
SafeHarbor: Defining Precise Decision Boundaries via Hierarchical Memory-Augmented Guardrail for LLM Agent Safety2605.05704v3 · Zhe Liu, Zonghao Ying, Wenxin Zhang et al.2026 · 0 citationsarXiv
TrajShield: Trajectory-Level Safety Mediation for Defending Text-to-Video Models Against Jailbreak Attacks2605.01761v1 · Quanchen Zou, Nizhang Li, Wenxin Zhang et al.2026 · 0 citationsarXiv
AgentVisor: Defending LLM Agents Against Prompt Injection via Semantic Virtualization2604.24118v1 · Zonghao Ying, Haozheng Wang, Jiangfan Liu et al.2026 · 0 citationsarXiv
Reading Between the Pixels: An Inscriptive Jailbreak Attack on Text-to-Image Models2604.05853v3 · Zonghao Ying, Haowen Dai, Lianyu Hu et al.2026 · 0 citationsarXiv
Uncovering Security Threats and Architecting Defenses in Autonomous Agents: A Case Study of OpenClaw2603.12644v1 · Zonghao Ying, Xiao Yang, Siyang Wu et al.2026 · 0 citationsarXiv
Reasoning-Oriented Programming: Chaining Semantic Gadgets to Jailbreak Large Vision Language Models2603.09246v1 · Quanchen Zou, Moyang Chen, Zonghao Ying et al.2026 · 0 citationsarXiv
Two Frames Matter: A Temporal Attack for Text-to-Video Model Jailbreaking2603.07028v1 · Moyang Chen, Zonghao Ying, Wenzhuo Xu et al.2026 · 0 citationsarXiv
Evolving Deception: When Agents Evolve, Deception Wins2603.05872v2 · Zonghao Ying, Haowen Dai, Tianyuan Zhang et al.2026 · 0 citationsarXiv
DIVER: Dynamic Iterative Visual Evidence Reasoning for Multimodal Fake News Detection2601.07178v1 · Weilin Zhou, Zonghao Ying, Chunlei Meng et al.2026 · 0 citationsarXiv
RoboSafe: Safeguarding Embodied Agents via Executable Safety Logic2512.21220v2 · Le Wang, Zonghao Ying, Xiao Yang et al.2025 · 0 citationsarXiv
Disentangling Fact from Sentiment: A Dynamic Conflict-Consensus Framework for Multimodal Fake News Detection2512.20670v2 · Weilin Zhou, Zonghao Ying, Rongchen Zhao et al.2025 · 0 citationsarXiv
SPARK: Jailbreaking T2V Models by Synergistically Prompting Auditory and Recontextualized Knowledge2511.13127v3 · Zonghao Ying, Moyang Chen, Nizhang Li et al.2025 · 0 citationsarXiv
Uncovering Strategic Egoism Behaviors in Large Language Models2511.09920v2 · Yaoyuan Zhang, Aishan Liu, Zonghao Ying et al.2025 · 0 citationsarXiv
Sequential Comics for Jailbreaking Multimodal Large Language Models via Structured Visual Storytelling2510.15068v1 · Deyue Zhang, Dongdong Yang, Junjie Mu et al.2025 · 0 citationsarXiv
SecureWebArena: A Holistic Security Evaluation Benchmark for LVLM-based Web Agents2510.10073v2 · Zonghao Ying, Yangguang Shao, Jianle Gan et al.2025 · 1 citationarXiv
Detoxifying Large Language Models via Autoregressive Reward Guided Representation Editing2510.01243v1 · Yisong Xiao, Aishan Liu, Siyuan Liang et al.2025 · 0 citationsarXiv
Mask-GCG: Are All Tokens in Adversarial Suffixes Necessary for Jailbreak Attacks?2509.06350v2 · Junjie Mu, Zonghao Ying, Zhekui Fan et al.2025 · 0 citations2026 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), pp. 13887-13891, 2026
PRISM: Programmatic Reasoning with Image Sequence Manipulation for LVLM Jailbreaking2507.21540v3 · Quanchen Zou, Zonghao Ying, Moyang Chen et al.2025 · 0 citationsarXiv
AGENTSAFE: Benchmarking the Safety of Embodied Agents on Hazardous Instructions2506.14697v3 · Zonghao Ying, Le Wang, Yisong Xiao et al.2025 · 0 citationsarXiv
Pushing the Limits of Safety: A Technical Report on the ATLAS Challenge 20252506.12430v2 · Zonghao Ying, Siyang Wu, Run Hao et al.2025 · 0 citationsarXiv
PRJ: Perception-Retrieval-Judgement for Generated Images2506.03683v1 · Qiang Fu, Zonglei Jing, Zonghao Ying et al.2025 · 0 citationsarXiv
Manipulating Multimodal Agents via Cross-Modal Prompt Injection2504.14348v4 · Le Wang, Zonghao Ying, Tianyuan Zhang et al.2025 · 8 citationsarXiv
Towards Understanding the Safety Boundaries of DeepSeek Models: Evaluation and Findings2503.15092v1 · Zonghao Ying, Guangyi Zheng, Yongxin Huang et al.2025 · 0 citationsarXiv
Probabilistic Modeling of Jailbreak on Multimodal LLMs: From Quantification to Application2503.06989v5 · Wenzhuo Xu, Zhipeng Wei, Xiongtao Sun et al.2025 · 0 citationsarXiv
Reasoning-Augmented Conversation for Multi-Turn Jailbreak Attacks on Large Language Models2502.11054v4 · Zonghao Ying, Deyue Zhang, Zonglei Jing et al.2025 · 6 citationsarXiv
CogMorph: Cognitive Morphing Attacks for Text-to-Image Models2501.11815v2 · Zonglei Jing, Zonghao Ying, Le Wang et al.2025 · 0 citationsarXiv
CS-Eval: A Comprehensive Large Language Model Benchmark for CyberSecurity2411.16239v3 · Zhengmin Yu, Jiutian Zeng, Siyi Chen et al.2024 · 0 citationsarXiv
SafeBench: A Safety Evaluation Framework for Multimodal Large Language Models2410.18927v1 · Zonghao Ying, Aishan Liu, Siyuan Liang et al.2024 · 17 citationsarXiv
DLP: towards active defense against backdoor attacks with decoupled learning process2406.13098v1 · Zonghao Ying, Bin Wu2024 · 2 citationsarXiv
NBA: defensive distillation for backdoor removal via neural behavior alignment2406.10846v1 · Zonghao Ying, Bin Wu2024 · 6 citationsarXiv
Unveiling the Safety of GPT-4o: An Empirical Study using Jailbreak Attacks2406.06302v2 · Zonghao Ying, Aishan Liu, Xianglong Liu et al.2024 · 2 citationsarXiv
Career total: 55 works. 51 are in this corpus.Showing the 50 most recent.
Profile built from the corpus for this byline.
Author records are still filling in while the Hub is in alpha. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.