TY  - RPRT
TI  - Aligning Large Vision-Language Models at Test Time: A Trajectory-Guided Structured Sampling Approach
AU  - Tianbao Jiang
AU  - Weicong Ni
AU  - Gerard de Melo
AU  - Linlin Wang
PY  - 2026
UR  - https://arxiv.org/abs/2608.03204
ID  - 2608.03204
ER  -