Coarse-to-Control: Action-Token Planning for Vision-Language-Action Models figure
AlphaXiv 中文概览(可滚动查看)