@article{vlmsneedwordsvisionlanguagemodelsignorev,
  title = {VLMs Need Words: Vision Language Models Ignore Visual Detail In Favor of Semantic Anchors},
  author = {Haz Sameen Shahgir and Xiaofu Chen and Yu Fu and Erfan Shayegani and Nael Abu-Ghazaleh and Yova Kementchedjhieva and Yue Dong},
  year = {2026},
  eprint = {2604.02486},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2604.02486},
}