5 papers
Differentially Private Natural Gradient Descent
Pan Li, Kai Chen, Shuai Chang +3
Under a fixed privacy budget, the utility of differentially private (DP) training is ultimately determined by its optimization efficiency. Standard first-order DP optimizers such a…
Turn-Based Structural Triggers: Structure-Conditioned Backdoors in Multi-Turn LLMs
Yiyang Lu, Jinwen He, Yue Zhao +4
Large Language Models (LLMs) are increasingly deployed as multi-turn assistants and customized through instruction tuning with project-specific training components. This practice c…
PrivacyXray: Detecting Privacy Breaches in LLMs through Semantic Consistency and Probability Certainty
Jinwen He, Yiyang Lu, Zijin Lin +2
Large Language Models (LLMs) are widely used in sensitive domains, including healthcare, finance, and legal services, raising concerns about potential private information leaks dur…
LLM Factoscope: Uncovering LLMs' Factual Discernment through Inner States Analysis
Jinwen He, Yujia Gong, Kai Chen +3
Large Language Models (LLMs) have revolutionized various domains with extensive knowledge and creative capabilities. However, a critical issue with LLMs is their tendency to produc…
I Don't Know You, But I Can Catch You: Real-Time Defense against Diverse Adversarial Patches for Object Detectors
Zijin Lin, Yue Zhao, Kai Chen +1
Deep neural networks (DNNs) have revolutionized the field of computer vision like object detection with their unparalleled performance. However, existing research has shown that DN…