Wang, J., Wu, Z., Ren, D., Han, Y., Zhao, W., & Li, W. (2026). TReVS: Integrating Textual Relevance and Visual Saliency for Efficient Vision-Language Model Token Pruning. https://omanscience.com/ar/articles/trevs-integrating-textual-relevance-and-visual-saliency-for-efficient-vision-language-model-token-pruning