CAC-VLA: Context-Gated Action Conditioning for Vision-Language-Action Models figure
AlphaXiv 中文概览(可滚动查看)