@misc{indiciae0edc0072d396,
  title = {Aligning Large Vision-Language Models at Test Time: A Trajectory-Guided Structured Sampling Approach},
  author = {Tianbao Jiang and Weicong Ni and Gerard de Melo and Linlin Wang},
  year = {2026},
  url = {https://arxiv.org/abs/2608.03204},
  note = {Source identifier: 2608.03204}
}