The Visual Iconicity Challenge: Evaluating Vision–Language Models on Sign Language Form–Meaning Mapping
Published in Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (ACL 2026, Main), 2026
We introduce the Visual Iconicity Challenge, a benchmark for evaluating whether vision–language models can map iconic form–meaning relationships in sign languages, revealing significant limitations in current VLMs.
Download here
