@article{dovisionlanguagemodelsreallyunderstand,
  title = {Do Vision-Language Models Really Understand Visual Language?},
  author = {Yifan Hou and Buse Giledereli and Yilei Tu and Mrinmaya Sachan},
  year = {2024},
  eprint = {2410.00193},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2410.00193v2},
}