2 papers
cs.LG2026
A Single Suffix to Break Them All: Basin-Aware Jailbreaks for Merged Model Families
Yu Zhe, Yixin Tan, Junhao Wei +1
Model merging enables combining multiple fine-tuned models without additional training, but its safety implications remain poorly understood. Prior work primarily attributes mergin…
cs.CV2026
Filtering Memorization from Parameter-Space in Diffusion Models
Yu Zhe, Yang Jiayan, Wei Junhao +2
Low-Rank Adaptation (LoRA) has become a widely used mechanism for customizing diffusion models, enabling users to inject new visual concepts or styles through lightweight parameter…