13 papers
FlowErase-RL: Rethinking Concept Erasure as Reward Optimization in Flow Matching Models
Yi Sun, Zhiqi Zhang, Xinhao Zhong +5
Recent advances in flow matching models have significantly improved text-to-image generation quality, but also introduce growing safety risks due to the generation of harmful or un…
Towards Distillation-Resistant Large Language Models: An Information-Theoretic Perspective
Hao Fang, Tianyi Zhang, Tianqu Zhuang +6
Proprietary large language models (LLMs) embody substantial economic value and are generally exposed only as black-box APIs, yet adversaries can still exploit their outputs to extr…
Emergent Hierarchical Structure in Large Language Models: An Information-Theoretic Framework for Multi-Scale Representation
Yukin Zhang, Qi Dong, Kemu Xu
Why do language models from different architecture families respond so differently to the same perturbation? We argue that the answer is not scale, but \emph{how architecture shape…
Enhancing Gradient Inversion Attacks in Federated Learning via Hierarchical Feature Optimization
Hao Fang, Wenbo Yu, Bin Chen +4
Federated Learning (FL) has emerged as a compelling paradigm for privacy-preserving distributed machine learning, allowing multiple clients to collaboratively train a global model…
Closing the Safety Gap: Surgical Concept Erasure in Visual Autoregressive Models
Xinhao Zhong, Yimin Zhou, Zhiqi Zhang +6
The rapid progress of visual autoregressive (VAR) models has brought new opportunities for text-to-image generation, but also heightened safety concerns. Existing concept erasure t…
One Perturbation is Enough: On Generating Universal Adversarial Perturbations against Vision-Language Pre-training Models
Hao Fang, Jiawei Kong, Wenbo Yu +5
Vision-Language Pre-training (VLP) models have exhibited unprecedented capability in many applications by taking full advantage of the multimodal alignment. However, previous studi…