6 citations · 7 across the 17 of their papers we have counts for
13 papers · 1 filter
ContextLeak: Exfiltrating LLM Agent Context via Malicious Tools
Yuqi Jia, Ruiqi Wang, Patrick Li +3
Exfiltrating an LLM agent's runtime context -- such as the user prompt, execution trajectory, and tool list -- poses severe security and privacy risks to users. Such attacks can be…
Dynamic Malicious Skills in Agentic AI
Tianhao Chen, Zhengyuan Jiang, Yuepeng Hu +2
Skills are a key enabling component of agentic AI. While they enhance agents' capabilities, they also introduce new attack surfaces. In this work, we investigate one such attack su…
MalTool: Malicious Tool Attacks on LLM Agents
Yuepeng Hu, Yuqi Jia, Mengyuan Li +2
In a malicious tool attack, an attacker uploads a malicious tool to a distribution platform; once a user inadvertently installs the tool and the LLM agent selects it during task ex…
Fingerprinting LLMs via Prompt Injection
Yuepeng Hu, Zhengyuan Jiang, Mengyuan Li +4
Large language models (LLMs) are often modified after release through post-processing such as post-training or quantization, which makes it challenging to determine whether one mod…
VideoMarkBench: Benchmarking Robustness of Video Watermarking
Zhengyuan Jiang, Moyang Guo, Kecen Li +5
The rapid development of video generative models has led to a surge in highly realistic synthetic videos, raising ethical concerns related to disinformation and copyright infringem…
Jailbreaking Safeguarded Text-to-Image Models via Large Language Models
Zhengyuan Jiang, Yuepeng Hu, Yuchen Yang +2
Text-to-Image models may generate harmful content, such as pornographic images, particularly when unsafe prompts are submitted. To address this issue, safety filters are often adde…
SafeText: Safe Text-to-image Models via Aligning the Text Encoder
Yuepeng Hu, Zhengyuan Jiang, Neil Zhenqiang Gong
Text-to-image models can generate harmful images when presented with unsafe prompts, posing significant safety and societal risks. Alignment methods aim to modify these models to e…
AI-generated Image Detection: Passive or Watermark?
Moyang Guo, Yuepeng Hu, Zhengyuan Jiang +4
While text-to-image models offer numerous benefits, they also pose significant societal risks. Detecting AI-generated images is crucial for mitigating these risks. Detection method…
Certifiably Robust Image Watermark
Zhengyuan Jiang, Moyang Guo, Yuepeng Hu +2
Generative AI raises many societal concerns such as boosting disinformation and propaganda campaigns. Watermarking AI-generated content is a key technology to address these concern…
Refusing Safe Prompts for Multi-modal Large Language Models
Zedian Shao, Hongbin Liu, Yuepeng Hu +1
Multimodal large language models (MLLMs) have become the cornerstone of today's generative AI ecosystem, sparking intense competition among tech giants and startups. In particular,…
Stable Signature is Unstable: Removing Image Watermark from Diffusion Models
Yuepeng Hu, Zhengyuan Jiang, Moyang Guo +1
Watermark has been widely deployed by industry to detect AI-generated images. A recent watermarking framework called \emph{Stable Signature} (proposed by Meta) roots watermark into…
Watermark-based Attribution of AI-Generated Content
Zhengyuan Jiang, Moyang Guo, Yuepeng Hu +2
Several companies have deployed watermark-based detection to identify AI-generated content. However, attribution--the ability to trace back to the user of a generative AI (GenAI) s…
A Transfer Attack to Image Watermarks
Yuepeng Hu, Zhengyuan Jiang, Moyang Guo +1
Watermark has been widely deployed by industry to detect AI-generated images. The robustness of such watermark-based detector against evasion attacks in the white-box and black-box…