AbsVis - Benchmarking How Humans and Vision-Language Models "See" Abstract Concepts in Images.

Tarun Tater, Diego Frassinelli, Sabine Schulte im Walde

Browse the full EMNLP paper archive.