Reasoning Paths with Reference Objects Elicit Quantitative Spatial Reasoning in Large Vision-Language Models.

Yuan-Hong Liao, Rafid Mahmood, Sanja Fidler, David Acuna

Browse the full EMNLP paper archive.