Submitted by Polina Fedotova 279 Green-VLA: Staged Vision-Language-Action Model for Generalist Robots Sber Robotics Center 34 7