activity
20242026
collaborators

41 papers

cs.CV2026

IDDM: Identity-Decoupled Personalized Diffusion Models with a Tunable Privacy-Utility Trade-off

Linyan Dai, Xinwei Zhang, Haoyang Li +2

Personalized text-to-image diffusion models (e.g., DreamBooth, LoRA) enable users to synthesize high-fidelity avatars from a few reference photos for social expression. However, on…

cs.CR2026

Grounding-Driven Attack: Improving Encoder-based Adversarial Transferability against Large Vision-Language Models

Xinwei Zhang, Li Bai, Tianwei Zhang +5

Large vision-language models (LVLMs) have achieved impressive performance across multimodal tasks, but their reliance on visual inputs exposes them to adversarial threats. Encoder-…

cs.CL2026

How Much Do Large Language Model Cheat on Evaluation? Benchmarking Overestimation under the One-Time-Pad-Based Framework

Zi Liang, Liantong Yu, Shiyu Zhang +2

Overestimation in evaluating large language models (LLMs) has become an increasing concern. Due to the contamination of public benchmarks or imbalanced model training, LLMs may ach…

cs.LG2026

Decision Potential Surface: A Theoretical and Practical Approximation of Large Language Model Decision Boundary

Zi Liang, Zhiyao Wu, Haoyang Shang +5

Decision boundary, the subspace of inputs where a machine learning model assigns equal classification probabilities to two classes, is pivotal in revealing core model properties an…

cs.CR2026

On the Adversarial Robustness of Large Vision-Language Models under Visual Token Compression

Xinwei Zhang, Hangcheng Liu, Li Bai +4

Visual token compression is widely used to accelerate large vision-language models (LVLMs) by pruning or merging visual tokens, yet its adversarial robustness remains unexplored. W…

cs.CL2026

Unlearning Isn't Deletion: Investigating Reversibility of Machine Unlearning in LLMs

Xiaoyu Xu, Xiang Yue, Yang Liu +5

Unlearning in large language models (LLMs) aims to remove specified data, but its efficacy is typically assessed with task-level metrics like accuracy and perplexity. We show that…