Submitted by Polina Fedotova 303 Green-VLA: Staged Vision-Language-Action Model for Generalist Robots Sber Robotics Center 37 7