@misc{indiciae0421216a7f67,
  title = {Learning to See Before Seeing: Demystifying LLM Visual Priors from Language Pre-training},
  author = {Junlin Han and Shengbang Tong and David Fan and Yufan Ren and Koustuv Sinha and Philip Torr and Filippos Kokkinos},
  year = {2025},
  url = {https://arxiv.org/abs/2509.26625},
  note = {Source identifier: 2509.26625}
}