Submitted by Polina Fedotova 325 Green-VLA: Staged Vision-Language-Action Model for Generalist Robots Sber Robotics Center 130 8