*, GigaBrain-0.5M\*: a VLA That Learns From World Model-Based Reinforcement Learning figure
AlphaXiv 中文概览(可滚动查看)