Submitted by Polina Fedotova 197 Green-VLA: Staged Vision-Language-Action Model for Generalist Robots Sber Robotics Center 29 4