4 citations · 20 across the 34 of their papers we have counts for
22 papers · 1 filter
Inverting the Hidden: Unveiling Multimodal Privacy Leakage in Collaborative LVLM Inference
Shuaifan Jin, Zhibo Wang, Qiyuan Wang +5
Collaborative inference deploys Large Vision-Language Models (LVLMs) by partitioning computation between edge devices and the cloud. While withholding raw inputs supposedly ensures…
From Role Prompt to Infinite Thinking: Exploiting Persona Conditioning for Inference Cost Attacks in LLMs
Zhiyi Mou, Wangze Ni, Tianfang Xiao +6
LLMs are increasingly deployed in real-world applications, making inference efficiency and service reliability critical concerns due to their substantial computational costs. Howev…
LoRA-Key: User-Centric LoRA Watermarking for Text-to-Image Diffusion Models
Yaopeng Wang, Qingliang Wang, Zhibo Wang +5
Low-Rank Adaptation (LoRA) has become a widely used mechanism for customizing text-to-image diffusion models, enabling lightweight modules that are shared, reused, and commercializ…
LoopTrap: Termination Poisoning Attacks on LLM Agents
Huiyu Xu, Zhibo Wang, Wenhui Zhang +4
Modern LLM agents solve complex tasks by operating in iterative execution loops, where they repeatedly reason, act, and self-evaluate progress to determine when a task is complete.…
RouteScan: A Non-Intrusive Approach to Auditing MoE LLMs Safety via Expert Routing Telemetry
Bo Lv, Zhiheng Xu, KeDong Xiu +4
As Mixture-of-Experts (MoE) architectures are increasingly adopted for scaling Large Language Models (LLMs), safety auditing becomes necessary to verify whether these models produc…
DualSentinel: A Lightweight Framework for Detecting Targeted Attacks in Black-box LLM via Dual Entropy Lull Pattern
Xiaoyi Pang, Xuanyi Hao, Pengyu Liu +3
Recent intelligent systems integrate powerful Large Language Models (LLMs) through APIs, but their trustworthiness may be critically undermined by targeted attacks like backdoor an…
RerouteGuard: Understanding and Mitigating Adversarial Risks for LLM Routing
Wenhui Zhang, Huiyu Xu, Zhibo Wang +4
Recent advancements in multi-model AI systems have leveraged LLM routers to reduce computational cost while maintaining response quality by assigning queries to the most appropriat…
Attack-Resistant Watermarking for AIGC Image Forensics via Diffusion-based Semantic Deflection
Qingyu Liu, Yitao Zhang, Zhongjie Ba +4
Protecting the copyright of user-generated AI images is an emerging challenge as AIGC becomes pervasive in creative workflows. Existing watermarking methods (1) remain vulnerable t…
PromptCOS: Towards Content-only System Prompt Copyright Auditing for LLMs
Yuchen Yang, Yiming Li, Hongwei Yao +6
System prompts are critical for shaping the behavior and output quality of large language model (LLM)-based applications, driving substantial investment in optimizing high-quality…
PT-Mark: Invisible Watermarking for Text-to-image Diffusion Models via Semantic-aware Pivotal Tuning
Yaopeng Wang, Huiyu Xu, Zhibo Wang +5
Watermarking for diffusion images has drawn considerable attention due to the widespread use of text-to-image diffusion models and the increasing need for their copyright protectio…
Interpretable LLM Guardrails via Sparse Representation Steering
Zeqing He, Zhibo Wang, Huiyu Xu +3
Large language models (LLMs) exhibit impressive capabilities in generation tasks but are prone to producing harmful, misleading, or biased content, posing significant ethical and s…
Privacy-Preserving LLM Embedding Transmission for End-Cloud Collaboration
Shuaifan Jin, Xiaoyi Pang, Zhibo Wang +4
Recent studies improve on-device language model (LM) inference through end-cloud collaboration, where the end device retrieves useful information from cloud databases to enhance lo…
Can Small Language Models Reliably Resist Jailbreak Attacks? A Comprehensive Evaluation
Zhibo Wang, Wenhui Zhang, Huiyu Xu +3
Small language models (SLMs) have emerged as promising alternatives to large language models (LLMs) due to their low computational demands, enhanced privacy guarantees, and compara…
JailbreakLens: Interpreting Jailbreak Mechanism in the Lens of Representation and Circuit
Zeqing He, Zhibo Wang, Zhixuan Chu +4
Despite the outstanding performance of Large language Models (LLMs) in diverse tasks, they are vulnerable to jailbreak attacks, wherein adversarial prompts are crafted to bypass th…
PointNCBW: Towards Dataset Ownership Verification for Point Clouds via Negative Clean-label Backdoor Watermark
Cheng Wei, Yang Wang, Kuofeng Gao +4
Recently, point clouds have been widely used in computer vision, whereas their collection is time-consuming and expensive. As such, point cloud datasets are the valuable intellectu…
RedAgent: Red Teaming Large Language Models with Context-aware Autonomous Language Agent
Huiyu Xu, Wenhui Zhang, Zhibo Wang +5
Recently, advanced Large Language Models (LLMs) such as GPT-4 have been integrated into many real-world applications like Code Copilot. These applications have significantly expand…
Breaking Secure Aggregation: Label Leakage from Aggregated Gradients in Federated Learning
Zhibo Wang, Zhiwei Chang, Jiahui Hu +4
Federated Learning (FL) exhibits privacy vulnerabilities under gradient inversion attacks (GIAs), which can extract private information from individual gradients. To enhance privac…
Textual Unlearning Gives a False Sense of Unlearning
Jiacheng Du, Zhibo Wang, Jie Zhang +3
Language Models (LMs) are prone to ''memorizing'' training data, including substantial sensitive user information. To mitigate privacy risks and safeguard the right to be forgotten…
SoK: On Gradient Leakage in Federated Learning
Jiacheng Du, Jiahui Hu, Zhibo Wang +4
Federated learning (FL) facilitates collaborative model training among multiple clients without raw data exposure. However, recent studies have shown that clients' private training…
Towards Deep Learning Models Resistant to Transfer-based Adversarial Attacks via Data-centric Robust Learning
Yulong Yang, Chenhao Lin, Xiang Ji +5
Transfer-based adversarial attacks raise a severe threat to real-world deep learning systems since they do not require access to target models. Adversarial training (AT), which is…
A Dual Stealthy Backdoor: From Both Spatial and Frequency Perspectives
Yudong Gao, Honglong Chen, Peng Sun +3
Backdoor attacks pose serious security threats to deep neural networks (DNNs). Backdoored models make arbitrarily (targeted) incorrect predictions on inputs embedded with well-desi…
"Adversarial Examples" for Proof-of-Learning
Rui Zhang, Jian Liu, Yuan Ding +3
In S&P '21, Jia et al. proposed a new concept/mechanism named proof-of-learning (PoL), which allows a prover to demonstrate ownership of a machine learning model by proving integri…