CoRE-VLA: Towards Scalable and Robust Vision-Language-Action Modeling via Conditional Routing of Experts figure
AlphaXiv 中文概览(可滚动查看)