Attribute-Comprehension-of-VLMs We provide a benchmark for evaluating the attribute understanding capabilities of large vision-language models.