3 citations · 5 across the 29 of their papers we have counts for
33 papers
SteerProbe: Learning to Bypass Safety Steering in Vision-Language Models
Xinwei Zhang, Aoting Hu, Hangcheng Liu +3
Activation steering offers an inference-time defense for vision--language models (VLMs) by modifying intermediate representations without updating backbone parameters. However, pro…
DIBench: Benchmarking Decision Integrity of GUI-based Mobile Agents Under Deceptive Injections
Li Hu, Kanghua Mo, Yingbin Jin +2
As GUI-based mobile agents rapidly progress, rigorous safety evaluation of their autonomous decision-making in realistic app interfaces becomes increasingly critical. Existing benc…
Inferring Hidden User Models from the Behavior of Personalized LLM Agents
Haoyang Li, Yaxin Xiao, Qingqing Ye +2
Recent personalized LLM agents increasingly transform information retained in memory into compressed or structured representations, which we call user models, to guide later decisi…
The Shape of Ownership: Verifying LLM Provenance through Semantic Structures
Zhongrui Sun, Jiahao Chen, Oubo Ma +4
As large language models (LLMs) are increasingly redistributed, adapted, and served behind opaque APIs, model ownership can no longer be established reliably by inspecting model in…
CAITLYN: Can LLM Agents Autonomously Synthesize Defenses against Emerging Injection Attacks?
Zi Liang, Xiaoyu Xu, Yanyun Wang +3
Prompt injection attacks on Large Language Model (LLM) agents seek to introduce malicious instructions or content into external text sources retrieved by agents, forcing the underl…
Whispers in the Noise: Surrogate-Guided Concept Awakening via a Multi-Agent Framework
Mengyu Sun, Ziyuan Yang, Zunlong Zhou +3
Diffusion models (DMs) are widely used for text-to-image generation, but their strong generative capabilities also raise concerns about unsafe or undesirable content. Concept erasu…
Cross-Modal Backdoors in Multimodal Large Language Models
Runhe Wang, Li Bai, Haibo Hu +1
Developers increasingly construct multimodal large language models (MLLMs) by assembling pretrained components,introducing supply-chain attack surfaces.Existing security research p…
When Routine Chats Turn Toxic: Unintended Long-Term State Poisoning in Personalized Agents
Xiaoyu Xu, Minxin Du, Qipeng Xie +3
Personalized LLM agents maintain persistent cross-session state to support long-horizon collaboration. Yet, this persistence introduces a subtle but critical security vulnerability…
Robust Alignment: Harmonizing Clean Accuracy and Adversarial Robustness in Adversarial Training
Yanyun Wang, Qingqing Ye, Li Liu +2
Adversarial Training (AT) is one of the most effective methods for developing robust deep neural networks (DNNs). However, AT faces a trade-off problem between clean accuracy and a…
Argus: Reorchestrating Static Analysis via a Multi-Agent Ensemble for Full-Chain Security Vulnerability Detection
Zi Liang, Qipeng Xie, Jun He +7
Recent advancements in Large Language Models (LLMs) have sparked interest in their application to Static Application Security Testing (SAST), primarily due to their superior contex…
Machine Unlearning in Low-Dimensional Feature Subspace
Kun Fang, Qinghua Tao, Junxu Liu +5
Machine Unlearning (MU) aims at removing the influence of specific data from a pretrained model while preserving performance on the remaining data. In this work, a novel perspectiv…
LURE: Latent Space Unblocking for Multi-Concept Reawakening in Diffusion Models
Mengyu Sun, Ziyuan Yang, Andrew Beng Jin Teoh +3
Concept erasure aims to suppress sensitive content in diffusion models, but recent studies show that erased concepts can still be reawakened, revealing vulnerabilities in erasure m…
From Domains to Instances: Dual-Granularity Data Synthesis for LLM Unlearning
Xiaoyu Xu, Minxin Du, Zitong Li +6
Although machine unlearning is essential for removing private, harmful, or copyrighted content from LLMs, current benchmarks often fail to faithfully represent the true ``forgettin…
Class-feature Watermark: A Resilient Black-box Watermark Against Model Extraction Attacks
Yaxin Xiao, Qingqing Ye, Zi Liang +4
Machine learning models constitute valuable intellectual property, yet remain vulnerable to model extraction attacks (MEA), where adversaries replicate their functionality through…
SEBA: Sample-Efficient Black-Box Attacks on Visual Reinforcement Learning
Tairan Huang, Yulin Jin, Junxu Liu +2
Visual reinforcement learning has achieved remarkable progress in visual control and robotics, but its vulnerability to adversarial perturbations remains underexplored. Most existi…
Virus Infection Attack on LLMs: Your Poisoning Can Spread "VIA" Synthetic Data
Zi Liang, Qingqing Ye, Xuan Liu +3
Synthetic data refers to artificial samples generated by models. While it has been validated to significantly enhance the performance of large language models (LLMs) during trainin…
Toward Efficient Inference Attacks: Shadow Model Sharing via Mixture-of-Experts
Li Bai, Qingqing Ye, Xinwei Zhang +4
Machine learning models are often vulnerable to inference attacks that expose sensitive information from their training data. Shadow model technique is commonly employed in such at…
StatEval: A Comprehensive Benchmark for Large Language Models in Statistics
Yuchen Lu, Run Yang, Yichen Zhang +6
Despite rapid advances in large language models (LLMs), statistical reasoning remains underrepresented in existing LLM benchmarks, which often do not reflect the layered, proof-dri…
Decision Potential Surface: A Theoretical and Practical Approximation of Large Language Model Decision Boundary
Zi Liang, Zhiyao Wu, Haoyang Shang +5
Decision boundary, the subspace of inputs where a machine learning model assigns equal classification probabilities to two classes, is pivotal in revealing core model properties an…
FreeTalk:A plug-and-play and black-box defense against speech synthesis attacks
Yuwen Pu, Zhou Feng, Chunyi Zhou +4
Recently, speech assistant and speech verification have been used in many fields, which brings much benefit and convenience for us. However, when we enjoy these speech applications…
Reminiscence Attack on Residuals: Exploiting Approximate Machine Unlearning for Privacy
Yaxin Xiao, Qingqing Ye, Li Hu +5
Machine unlearning enables the removal of specific data from ML models to uphold the right to be forgotten. While approximate unlearning algorithms offer efficient alternatives to…
How Much Do Large Language Model Cheat on Evaluation? Benchmarking Overestimation under the One-Time-Pad-Based Framework
Zi Liang, Liantong Yu, Shiyu Zhang +2
Overestimation in evaluating large language models (LLMs) has become an increasing concern. Due to the contamination of public benchmarks or imbalanced model training, LLMs may ach…
United Minds or Isolated Agents? Exploring Coordination of LLMs under Cognitive Load Theory
HaoYang Shang, Xuan Liu, Zi Liang +3
Large Language Models (LLMs) exhibit a notable performance ceiling on complex, multi-faceted tasks. As practitioners increasingly rely on heavy context engineering -- curating intr…
Does Low Rank Adaptation Lead to Lower Robustness against Training-Time Attacks?
Zi Liang, Haibo Hu, Qingqing Ye +2
Low rank adaptation (LoRA) has emerged as a prominent technique for fine-tuning large language models (LLMs) thanks to its superb efficiency gains over previous methods. While exte…