@misc{indiciaef8857e8dbdb0,
  title = {IVRA: Improving Visual-Token Relations for Robot Action Policy with Training-Free Hint-Based Guidance},
  author = {Jongwoo Park and Kanchana Ranasinghe and Jinhyeok Jang and Cristina Mata and Yoo Sung Jang and Michael S Ryoo},
  year = {2026},
  url = {https://arxiv.org/abs/2601.16207},
  note = {Source identifier: 2601.16207}
}