Our work “Seeing Isn’t Believing: Uncovering Blind Spots in Evaluator Vision-Language Models” is out on arXiv!