12 citations · 25 across the 13 of their papers we have counts for
14 papers
Scenes as an Attack Surface: Uncovering Scene-Grounded Safety Failures in MLLMs
Tongxi Wu, Yuyang Cheng, Jiacheng Xian +5
Existing jailbreak attacks have exposed serious safety weaknesses in multimodal large language models (MLLMs). Most prior attacks rely on specially constructed visual inputs to car…
Furina: Fragmented Uncertainty-Driven Refusal Instability Attack
Tongxi Wu, Jian Zhang, Yang Gao
Safety alignment in large language models (LLMs) and multimodal large language models (MLLMs) is commonly assumed to operate as a near-binary threshold mechanism. We challenge this…
Can Small Training Runs Reliably Guide Data Curation? Rethinking Proxy-Model Practice
Jiachen T. Wang, Tong Wu, Kaifeng Lyu +4
Data teams at frontier AI companies routinely train small proxy models to make critical decisions about pretraining data recipes for full-scale training runs. However, the communit…
ReliabilityRAG: Effective and Provably Robust Defense for RAG-based Web-Search
Zeyu Shen, Basileal Imana, Tong Wu +3
Retrieval-Augmented Generation (RAG) enhances Large Language Models by grounding their outputs in external documents. These systems, however, remain vulnerable to attacks on the re…
Does More Inference-Time Compute Really Help Robustness?
Tong Wu, Chong Xiang, Jiachen T. Wang +4
Recently, Zaremba et al. demonstrated that increasing inference-time computation improves robustness in large proprietary reasoning LLMs. In this paper, we first show that smaller-…
Hiding in Plain Sight: A Steganographic Approach to Stealthy LLM Jailbreaks
Jianing Geng, Biao Yi, Zekun Fei +4
Jailbreak attacks pose a serious threat to Large Language Models (LLMs) by bypassing their safety mechanisms. A truly advanced jailbreak is defined not only by its effectiveness bu…
Effectively Controlling Reasoning Models through Thinking Intervention
Tong Wu, Chong Xiang, Jiachen T. Wang +2
Reasoning-enhanced large language models (LLMs) explicitly generate intermediate reasoning steps prior to generating final answers, helping the model excel in complex problem-solvi…
Instructional Segment Embedding: Improving LLM Safety with Instruction Hierarchy
Tong Wu, Shujian Zhang, Kaiqiang Song +7
Large Language Models (LLMs) are susceptible to security and safety threats, such as prompt injection, prompt extraction, and harmful requests. One major cause of these vulnerabili…
Certifiably Robust RAG against Retrieval Corruption
Chong Xiang, Tong Wu, Zexuan Zhong +3
Retrieval-augmented generation (RAG) is susceptible to retrieval corruption attacks, where malicious passages injected into retrieval results can lead to inaccurate model responses…
Position: Towards Resilience Against Adversarial Examples
Sihui Dai, Chong Xiang, Tong Wu +1
Current research on defending against adversarial examples focuses primarily on achieving robustness against a single attack type such as or -bounded attack…
PatchCURE: Improving Certifiable Robustness, Model Utility, and Computation Efficiency of Adversarial Patch Defenses
Chong Xiang, Tong Wu, Sihui Dai +3
State-of-the-art defenses against adversarial patch attacks can now achieve strong certifiable robustness with a marginal drop in model utility. However, this impressive performanc…
Privacy-Preserving In-Context Learning for Large Language Models
Tong Wu, Ashwinee Panda, Jiachen T. Wang +1
In-context learning (ICL) is an important capability of Large Language Models (LLMs), enabling these models to dynamically adapt based on specific, in-context exemplars, thereby im…
Just Rotate it: Deploying Backdoor Attacks via Rotation Transformation
Tong Wu, Tianhao Wang, Vikash Sehwag +2
Recent works have demonstrated that deep learning models are vulnerable to backdoor poisoning attacks, where these attacks instill spurious correlations to external trigger pattern…
Defending Against Physically Realizable Attacks on Image Classification
Tong Wu, Liang Tong, Yevgeniy Vorobeychik
We study the problem of defending deep neural network approaches for image classification from physically realizable attacks. First, we demonstrate that the two most scalable and e…