@misc{indiciae315f5c2bf913,
  title = {Improving Visual Token Reduction via Rectifying Distortions for Efficient Multimodal LLM Inference},
  author = {Hyeonwoo Cho and Donghyeon Baek and Yewon Kim and Bumsub Ham},
  year = {2026},
  url = {https://arxiv.org/abs/2606.01711},
  note = {Source identifier: 2606.01711}
}