REVISION: Rendering Tools Enable Spatial Fidelity in Vision-Language Models Paper • 2408.02231 • Published Aug 5 • 2