CVT-Bench: Probing Spatial-State Integrity through Counterfactual Viewpoint Transformations

Shanmukha Vellamcheti, Uday Kiran Kothapalli, Disharee Bhowmick +1

Multimodal large language models (MLLMs) perform strongly on isolated spatial tasks, but whether their predictions remain persistent and mutually coherent across viewpoints and com…