VLA-Adapter: An Effective Paradigm for Tiny-Scale Vision-Language-Action Model - Explained Simply | ArXiv Explained