@misc{indiciae756c5afbc9cc,
  title = {HAP: Head-Adaptive Visual Token Pruning via Cross-Modal Alignment},
  author = {Yuanhao Sun and Huawei Ji and Yuan Jin and Cheng Deng and Luoyi Fu and Xinbing Wang},
  year = {2026},
  url = {https://arxiv.org/abs/2608.23921},
  note = {Source identifier: 2608.23921}
}