@article{asapattentionshiftawarepruningforefficie,
  title = {ASAP: Attention-Shift-Aware Pruning for Efficient LVLM Inference},
  author = {Surendra Pathak and Bo Han},
  year = {2026},
  eprint = {2603.14549},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2603.14549},
}