Preprint
Sep 2026
Composition, Not Conversation: VLMs Lose the Scene, Not the Thread
Vision-language models (VLMs) increasingly reason over visual evidence that is cropped, segmented, retrieved, or revealed over time. Yet most VQA benchmarks present the complete image and question at once. We ask what models lose when the same information is fragmented. We introduce Layered-VQA, with 93 scenes and 300...
L. D. M. S. Sai Teja, Ufaq Khan, N. S. G. Krishna et al.
· 0 citations