Fetching the paper…

VisNumBench: Evaluating Number Sense of Multimodal Large Language Models · Around