Submitted by Polina Fedotova 323 Green-VLA: Staged Vision-Language-Action Model for Generalist Robots Sber Robotics Center 121 8