AntiGrounding: Lifting Robotic Actions into VLM Representation Space for Decision Making

Open in new window