G^3VLA: Geometric inductive bias for Vision-Language-Action Models Paper • 2606.24472 • Published about 1 month ago • 3