We would like to kindly request adding our recent paper:
[ICLR 2026] AgilePruner: An Empirical Study of Attention and Diversity for Adaptive Visual Token Pruning in Large Vision-Language Models
Paper: http://arxiv.org/abs/2603.01236
Code: https://github.com/cvsp-lab/AgilePruner
In this work, we provide an empirical analysis of visual token pruning in LVLMs using effective rank (erank) and attention entropy, revealing the strengths and limitations of attention- and diversity-based methods and offering insights for adaptive pruning design.
Thank you for your consideration.
We would like to kindly request adding our recent paper:
[ICLR 2026] AgilePruner: An Empirical Study of Attention and Diversity for Adaptive Visual Token Pruning in Large Vision-Language Models
Paper: http://arxiv.org/abs/2603.01236
Code: https://github.com/cvsp-lab/AgilePruner
In this work, we provide an empirical analysis of visual token pruning in LVLMs using effective rank (erank) and attention entropy, revealing the strengths and limitations of attention- and diversity-based methods and offering insights for adaptive pruning design.
Thank you for your consideration.