11 citations · 39 across the 26 of their papers we have counts for
27 papers
Finetune Like You Pretrain: Boosting Zero-shot Adversarial Robustness in Vision-language Models
Songlong Xing, Weijie Wang, Zhengyu Zhao +3
Despite their impressive zero-shot abilities, vision-language models such as CLIP have been shown to be susceptible to adversarial attacks. To enhance its adversarial robustness, r…
SkillJect: Effectively Automating Skill-Based Prompt Injection for Skill-Enabled Agents
Xiaojun Jia, Jie Liao, Simeng Qin +5
Agent skills extend LLM agents with task-specific instructions, executable scripts, and auxiliary resources, improving reusability but creating a new supply-chain attack surface. A…
Reimagining Safety Alignment with An Image
Yifan Xia, Guorui Chen, Wenqian Yu +3
Large language models (LLMs) excel in diverse applications but face dual challenges: generating harmful content under jailbreak attacks and over-refusal of benign queries due to ri…
Can Knowledge-Graph-based Retrieval Augmented Generation Really Retrieve What You Need?
Junchi Yu, Yujie Liu, Jindong Gu +2
Retrieval-Augmented Generation (RAG) based on knowledge graphs (KGs) enhances large language models (LLMs) by providing structured and interpretable external knowledge. However, ex…
Can an Individual Manipulate the Collective Decisions of Multi-Agents?
Fengyuan Liu, Rui Zhao, Shuo Chen +4
Individual Large Language Models (LLMs) have demonstrated significant capabilities across various domains, such as healthcare and law. Recent studies also show that coordinated mul…
LLM Jailbreak Detection for (Almost) Free!
Guorui Chen, Yifan Xia, Xiaojun Jia +3
Large language models (LLMs) enhance security through alignment when widely used, but remain susceptible to jailbreak attacks capable of producing inappropriate content. Jailbreak…
AlignGuard: Scalable Safety Alignment for Text-to-Image Generation
Runtao Liu, I Chieh Chen, Jindong Gu +6
Text-to-image (T2I) models are widespread, but their limited safety guardrails expose end users to harmful content and potentially allow for model misuse. Current safety measures a…
Benchmarking Open-ended Audio Dialogue Understanding for Large Audio-Language Models
Kuofeng Gao, Shu-Tao Xia, Ke Xu +2
Large Audio-Language Models (LALMs), such as GPT-4o, have recently unlocked audio dialogue capabilities, enabling direct spoken exchanges with humans. The potential of LALMs broade…
HTS-Attack: Heuristic Token Search for Jailbreaking Text-to-Image Models
Sensen Gao, Xiaojun Jia, Yihao Huang +5
Text-to-Image(T2I) models have achieved remarkable success in image generation and editing, yet these models still have many potential issues, particularly in generating inappropri…
Can Editing LLMs Inject Harm?
Canyu Chen, Baixiang Huang, Zekun Li +12
Large Language Models (LLMs) have emerged as a new information channel. Meanwhile, one critical but under-explored question is: Is it possible to bypass the safety alignment and in…
Learning Visual Prompts for Guiding the Attention of Vision Transformers
Razieh Rezaei, Masoud Jalili Sabet, Jindong Gu +3
Visual prompting infuses visual information into the input image to adapt models toward specific predictions and tasks. Recently, manually crafted markers such as red circles are s…
Improved Techniques for Optimization-Based Jailbreaking on Large Language Models
Xiaojun Jia, Tianyu Pang, Chao Du +5
Large language models (LLMs) are being rapidly developed, and a key component of their widespread deployment is their safety-related alignment. Many red-teaming efforts aim to jail…
Special Characters Attack: Toward Scalable Training Data Extraction From Large Language Models
Yang Bai, Ge Pei, Jindong Gu +2
Large language models (LLMs) have achieved remarkable performance on a wide range of tasks. However, recent studies have shown that LLMs can memorize training data and simple repea…
Energy-Latency Manipulation of Multi-modal Large Language Models via Verbose Samples
Kuofeng Gao, Jindong Gu, Yang Bai +4
Despite the exceptional performance of multi-modal large language models (MLLMs), their deployment requires substantial computational resources. Once malicious users induce high en…
Latent Guard: a Safety Framework for Text-to-image Generation
Runtao Liu, Ashkan Khakzar, Jindong Gu +3
With the ability to generate high-quality images, text-to-image (T2I) models can be exploited for creating inappropriate content. To prevent misuse, existing safety measures are ei…
As Firm As Their Foundations: Can open-sourced foundation models be used to create adversarial examples for downstream tasks?
Anjun Hu, Jindong Gu, Francesco Pinto +2
Foundation models pre-trained on web-scale vision-language data, such as CLIP, are widely used as cornerstones of powerful machine learning systems. While pre-training offers clear…
An Image Is Worth 1000 Lies: Adversarial Transferability across Prompts on Vision-Language Models
Haochen Luo, Jindong Gu, Fengyuan Liu +1
Different from traditional task-specific vision models, recent large VLMs can readily adapt to different vision tasks by simply using different textual instructions, i.e., prompts.…
Hide in Thicket: Generating Imperceptible and Rational Adversarial Perturbations on 3D Point Clouds
Tianrui Lou, Xiaojun Jia, Jindong Gu +4
Adversarial attack methods based on point manipulation for 3D point cloud classification have revealed the fragility of 3D models, yet the adversarial examples they produce are eas…
Inducing High Energy-Latency of Large Vision-Language Models with Verbose Images
Kuofeng Gao, Yang Bai, Jindong Gu +4
Large vision-language models (VLMs) such as GPT-4 have achieved exceptional performance across various multi-modal tasks. However, the deployment of VLMs necessitates substantial e…
Does Few-shot Learning Suffer from Backdoor Attacks?
Xinwei Liu, Xiaojun Jia, Jindong Gu +3
The field of few-shot learning (FSL) has shown promising results in scenarios where training data is limited, but its vulnerability to backdoor attacks remains largely unexplored.…
OT-Attack: Enhancing Adversarial Transferability of Vision-Language Models via Optimal Transport Optimization
Dongchen Han, Xiaojun Jia, Yang Bai +3
Vision-language pre-training (VLP) models demonstrate impressive abilities in processing both images and text. However, they are vulnerable to multi-modal adversarial examples (AEs…
TranSegPGD: Improving Transferability of Adversarial Examples on Semantic Segmentation
Xiaojun Jia, Jindong Gu, Yihao Huang +4
Transferability of adversarial examples on image classification has been systematically explored, which generates adversarial examples in black-box mode. However, the transferabili…
Initialization Matters for Adversarial Transfer Learning
Andong Hua, Jindong Gu, Zhiyu Xue +3
With the prevalence of the Pretraining-Finetuning paradigm in transfer learning, the robustness of downstream tasks has become a critical concern. In this work, we delve into adver…
Improving Adversarial Transferability via Model Alignment
Avery Ma, Amir-massoud Farahmand, Yangchen Pan +2
Neural networks are susceptible to adversarial perturbations that are transferable across different models. In this paper, we introduce a novel model alignment technique aimed at i…