What Vision-Language Models `See' when they See Scenes

VLM
Re-assign community