Fetching the paper…

ROSA: Harnessing Robot States for Vision-Language and Action Alignment · Around