GuidedVLA: Specifying Task-Relevant Factors via Plug-and-Play Action Attention Specialization figure
AlphaXiv 中文概览(可滚动查看)