1 paper · 1 filter
Nicholas Budny, Kia Ghods, Declan Campbell +6
Why do Vision Language Models (VLMs), despite success on standard benchmarks, often fail to match human performance on surprisingly simple visual reasoning tasks? While the underly…