VLM4D: Towards Spatiotemporal Awareness in Vision Language Models Paper • 2508.02095 • Published Aug 4 • 9 • 2