12 citations · 35 across the 42 of their papers we have counts for
51 papers
GeoCloak: An Agentic Framework for Semantics-Preserving Geolocation Privacy via Peripheral Clue Manipulation
Yixin Wu, Mingjie Li, Yun Shen +2
The rapid advancement of large vision language models (LVLMs) has made geolocation inference from images highly accurate and accessible. This poses serious privacy risks: a passive…
"Shut Up and Let Me Enjoy My Otome": Understanding and Measuring the Toxicity in Otome Game Communities
Yage Zhang, Xinyue Shen, Yukun Jiang +2
Otome games, a romance simulation genre primarily targeting female, have emerged as a major force in the global gaming market, attracting hundreds of millions of players and billio…
From Celebrities to Anyone: Characterizing AI Nudification Content, Technology, and Community Dynamics on 4chan
Chi Cui, Yixin Wu, Yang Zhang
AI nudification uses generative models to create synthetic non-consensual sexually explicit imagery (SNEACI) of real individuals. Prior work has examined dedicated nudification pla…
BadBone: Backdoor Attacks Against Backbone Models in Visual Prompt Learning
Ziqing Yang, Rui Wen, Xinlei He +3
Prompt learning is a new machine learning paradigm that has attracted ample attention due to its simplicity and proven efficacy. Despite its growing adoption, the security vulnerab…
A Systematic Study of Training-Free Methods for Trustworthy Large Language Models
Wai Man Si, Mingjie Li, Michael Backes +1
As Large Language Models (LLMs) receive increasing attention and are being deployed across various domains, their potential risks, including generating harmful or biased content, p…
Pruning Unsafe Tickets: A Resource-Efficient Framework for Safer and More Robust LLMs
Wai Man Si, Mingjie Li, Michael Backes +1
Machine learning models are increasingly deployed in real-world applications, but even aligned models such as Mistral and LLaVA still exhibit unsafe behaviors inherited from pre-tr…
HarmfulSkillBench: How Do Harmful Skills Weaponize Your Agents?
Yukun Jiang, Yage Zhang, Michael Backes +2
Large language models (LLMs) have evolved into autonomous agents that rely on open skill ecosystems (e.g., ClawHub and Skills.Rest), hosting numerous publicly reusable skills. Exis…
Triviality Corrected Endogenous Reward
Xinda Wang, Zhengxu Hou, Yangshijie Zhang +6
Reinforcement learning for open-ended text generation is constrained by the lack of verifiable rewards, necessitating reliance on judge models that require either annotated data or…
Seeing but Not Thinking: Routing Distraction in Multimodal Mixture-of-Experts
Haolei Xu, Haiwen Hong, Hongxing Li +7
Multimodal Mixture-of-Experts (MoE) models have achieved remarkable performance on vision-language tasks. However, we identify a puzzling phenomenon termed Seeing but Not Thinking:…
The Art of (Mis)alignment: How Fine-Tuning Methods Effectively Misalign and Realign LLMs in Post-Training
Rui Zhang, Hongwei Li, Yun Shen +6
The deployment of large language models (LLMs) raises significant ethical and safety concerns. While LLM alignment techniques are adopted to improve model safety and trustworthines…
CT Open: An Open-Access, Uncontaminated, Live Platform for the Open Challenge of Clinical Trial Outcome Prediction
Jianyou Wang, Youze Zheng, Longtian Bao +11
Scientists have long sought to accurately predict outcomes of real-world events before they happen. Can AI systems do so more reliably? We study this question through clinical tria…
When Safe Concepts Become Unsafe: Multi-Concept Compositional Vulnerabilities in Text-to-Image Models
Chaoshuo Zhang, Yibo Liang, Mengke Tian +7
Text-to-image (T2I) models are increasingly optimized for following user instructions faithfully. However, we find that this capability introduces a safety vulnerability we call Mu…
When Understanding Becomes a Risk: Authenticity and Safety Risks in the Emerging Image Generation Paradigm
Ye Leng, Junjie Chu, Mingjie Li +5
Recently, multimodal large language models (MLLMs) have emerged as a unified paradigm for language and image generation. Compared with diffusion models, MLLMs possess a much strong…
AdaQE-CG: Adaptive Query Expansion for Web-Scale Generative AI Model and Data Card Generation
Haoxuan Zhang, Ruochi Li, Zhenni Liang +7
Transparent and standardized documentation is essential for building trustworthy generative AI (GAI) systems. However, existing automated methods for generating model and data card…
Finding and Reactivating Post-Trained LLMs' Hidden Safety Mechanisms
Mingjie Li, Wai Man Si, Michael Backes +2
Despite the impressive performance of general-purpose large language models (LLMs), they often require fine-tuning or post-training to excel at specific tasks. For instance, large…
Benchmark of Benchmarks: Unpacking Influence and Code Repository Quality in LLM Safety Benchmarks
Junjie Chu, Xinyue Shen, Ye Leng +3
The rapid expansion of research in LLM safety presents challenges in tracking advancements, making benchmarks important evaluation infrastructures for identifying key trends and fa…
Sparse Models, Sparse Safety: Unsafe Routes in Mixture-of-Experts LLMs
Yukun Jiang, Hai Huang, Mingjie Li +3
By introducing routers to selectively activate experts in Transformer layers, the mixture-of-experts (MoE) architecture significantly reduces computational costs in large language…
"Humans welcome to observe": A First Look at the Agent Social Network Moltbook
Yukun Jiang, Yage Zhang, Xinyue Shen +2
The rapid advancement of artificial intelligence (AI) agents has catalyzed the transition from static language models to autonomous agents capable of tool use, long-term planning,…
GEO-Detective: Unveiling Location Privacy Risks in Images with LLM Agents
Xinyu Zhang, Yixin Wu, Boyang Zhang +4
Images shared on social media often expose geographic cues. While early geolocation methods required expert effort and lacked generalization, the rise of Large Vision Language Mode…
AttackPilot: Autonomous Inference Attacks Against ML Services With LLM-Based Agents
Yixin Wu, Rui Wen, Chi Cui +2
Inference attacks have been widely studied and offer a systematic risk assessment of ML services; however, their implementation and the attack parameters for optimal estimation are…
From Evidence to Verdict: An Agent-Based Forensic Framework for AI-Generated Image Detection
Mengfei Liang, Yiting Qu, Yukun Jiang +2
The rapid evolution of AI-generated images poses growing challenges to information integrity and media authenticity. Existing detection approaches face limitations in robustness, i…
Adjacent Words, Divergent Intents: Jailbreaking Large Language Models via Task Concurrency
Yukun Jiang, Mingjie Li, Michael Backes +1
Despite their superior performance on a wide range of domains, large language models (LLMs) remain vulnerable to misuse for generating harmful content, a risk that has been further…
Amplifying Machine Learning Attacks Through Strategic Compositions
Yugeng Liu, Zheng Li, Hai Huang +2
Machine learning (ML) models are proving to be vulnerable to a variety of attacks that allow the adversary to learn sensitive information, cause mispredictions, and more. While the…
Excessive Reasoning Attack on Reasoning LLMs
Wai Man Si, Mingjie Li, Michael Backes +1
Recent reasoning large language models (LLMs), such as OpenAI o1 and DeepSeek-R1, exhibit strong performance on complex tasks through test-time inference scaling. However, prior st…