Elamur Research⌕
☰
home / papers / 2026 / Reducing Token Redundancy in LVLMs: A Systematic Review of Token Pruning Methods

Reducing Token Redundancy in LVLMs: A Systematic Review of Token Pruning Methods

Paper · 2026 · Hanzhang Yuan, Mengxuan Hu, Wenhao Zhang et al.

Summary

Modern LVLMs turn high-resolution images, multi-image inputs and long videos into very long sequences of visual tokens. Because self-attention in the prefill stage grows quadratically with sequence length, these tokens become the main cost driver. Many of them add little to the answer for a given prompt. This survey covers visual token pruning methods published since 2024 and claims to be the first comprehensive review focused on this topic.

The authors sort methods by where pruning occurs and which signals guide it:

  • Vision-side: inside or right after the ViT encoder, before the prefill stage.
  • LLM-side: inside the decoder, using cross-modal cues.
  • Hybrid: combinations of both.

Within these groups they compare importance criteria (similarity, semantics, attention, learned scores) and one-shot versus progressive application. The paper also separates pruning from neighbouring techniques such as merging, KV-cache eviction and weight pruning, and reviews evaluation protocols. It then offers practical guidance for choosing a method and outlines open problems, including video redundancy, hallucination under heavy pruning, and agentic tasks. The result is a shared structure for comparing work in a fast-growing efficiency area.

Key points

  • Taxonomy:
    • Vision-side pruning is split into similarity-based (LLaVA-PruMerge, DivPrune), semantic-aware (CDPruner, MFPruner, ALTP, MMG-Vid) and attention-based (FoPru, HiRED, VisPruner, HoloV, PTP).
    • LLM-side pruning covers progressive attention-based methods (FastV, PyramidDrop, SparseVLM, FitPrune), one-shot methods (DART, GreedyPrune) and multi-factor methods (V2Drop, FrameFusion, MoB).
    • Hybrid examples include ToDRE, MustDrop, VisionDrop and Vscan.
  • Why vision-side reduction pays off most: the prefill stage dominates cost, so removing tokens before or during prefill saves far more than pruning only at decoding time.
  • Three-step decision guide:
    • Step 1: identify the bottleneck (prefill versus KV-cache or decoding).
    • Step 2: match the task. General VQA and captioning tolerate roughly 70–80% pruning, while OCR, document QA and grounding call for ≤40%.
    • Step 3: check deployment constraints, such as whether training or extra parameters are allowed.
  • Encoder dependence: methods that rely on the [CLS] token (e.g., LLaVA-PruMerge, HiRED) assume ViT-based encoders and cannot be applied directly to models like Qwen2.5-VL.
  • Open problems and limitations:
    • The field lacks pruning-specific benchmarks and needs better ways to tell real retained information apart from hallucination at high pruning rates.
    • Hour-long video and agentic reasoning remain open challenges.
    • The review mostly covers late-fusion, ViT-based architectures.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓University of VirginiaCharlottesville · United States40 works · 10,099 citations🏢AppleCupertino · United States23 works · 2,318 citations

Also on the byline: Merck & Co., Inc., Rahway, NJ, USA.

 Fields, topics and concepts

📚 Shelves 5The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksArts and Humanities · 17,450 worksLanguage and Linguistics · 17,296 worksComputer Vision and Pattern Recognition · 3,486 works

🧭 Topics 1Research subjects and languages the work is about.Efficient models & compression · 297 works

Cite

Cite as Hanzhang Yuan, Mengxuan Hu, Wenhao Zhang, Tianlong Wang, Zhongliang Zhou, Jiasen Lu and Sheng Li. 2026. Reducing Token Redundancy in LVLMs: A Systematic Review of Token Pruning Methods. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-long.328

@inproceedings{yuan2026reducing,
  title = {{Reducing Token Redundancy in LVLMs: A Systematic Review of Token Pruning Methods}},
  author = {Hanzhang Yuan, Mengxuan Hu, Wenhao Zhang, Tianlong Wang, Zhongliang Zhou, Jiasen Lu and Sheng Li},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.328},
  url = {https://doi.org/10.18653/v1/2026.acl-long.328}
}

References

99 references, 29 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Mohammad Mahdi Abootorabi, Amirhosein Zobeiri, Mahdi Dehghani, Mohammadali Mohammadkhani, Bardia Mohammadi, Omid Ghahroodi, Mahdieh Soleymani Baghshah, and Ehsaneddin Asgari. 2025. Ask in any modality: A comprehensive survey on multimodal retrieval-augmented generation. Preprint, arXiv:2502.08826. find
  2. Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, and 1 others. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774.
  3. Saeed Ranjbar Alvar, Gursimran Singh, Mohammad Akbari, and Yong Zhang. 2025. Divprune: Diversity-based visual token pruning for large multimodal models. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 9392–9401. find
  4. Kazi Hasan Ibn Arif, JinYi Yoon, Dimitrios S Nikolopoulos, Hans Vandierendonck, Deepu John, and Bo Ji. 2025. Hired: Attention-guided token dropping for efficient inference of high-resolution vision-language models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 17… find
  5. Bizhe Bai, Jianjian Cao, Yadan Luo, and Tao Chen. 2025. Local information matters: Inference acceleration for grounded conversation generation models through adaptive local-aware token pruning. Preprint, arXiv:2503.23959. — [arXiv:2503.23959](https://arxiv.org/abs/2503.23959) find
  6. Daniel Bolya, Cheng-Yang Fu, Xiaoliang Dai, Peizhao Zhang, Christoph Feichtenhofer, and Judy Hoffman. 2023. Token merging: Your vit but faster. Preprint, arXiv:2210.09461. — [arXiv:2210.09461](https://arxiv.org/abs/2210.09461) find
  7. Adrian Bulat, Yassine Ouali, and Georgios Tzimiropoulos. 2025. Fwd2bot: Lvlm visual token compression with double forward bottleneck. Preprint, arXiv:2503.21757. — [arXiv:2503.21757](https://arxiv.org/abs/2503.21757) find
  8. Hanning Chen, Yang Ni, Wenjun Huang, Hyunwoo Oh, Yezi Liu, Tamoghno Das, and Mohsen Imani. 2025a. Lvlm_csp: Accelerating large vision language models via clustering, scattering, and pruning for reasoning segmentation. Preprint, arXiv:2504.10854. — [arXiv:2504.10854](https://arxiv.org/abs/2504.10854) find
  9. Junjie Chen, Xuyang Liu, Zichen Wen, Yiyu Wang, Siteng Huang, and Honggang Chen. 2025b. Variation-aware vision token dropping for faster large vision-language models. Preprint, arXiv:2509.01552. — [arXiv:2509.01552](https://arxiv.org/abs/2509.01552) find
  10. Liang Chen, Haozhe Zhao, Tianyu Liu, Shuai Bai, Junyang Lin, Chang Zhou, and Baobao Chang. 2024. An image is worth 1/2 tokens after layer 2: Plug-and-play inference acceleration for large vision-language models. In European Conference on Computer Vision, pages 19–35. Springer.
  11. Jian Cheng, Haidong Kang, Yuxin Shao, Nan Li, Pengjun Chen, Rui Wang, Saiqin Long, Xiaochun Yang, and Lianbo Ma. 2025. Survey on efficient large language models: Principles, algorithms, applications, and open issues. IEEE Transactions on Neural Networks and Learning Systems, pages 1–21. find
  12. Tzu-Chun Chien, Chieh-Kai Lin, Shiang-Feng Tsai, Ruei-Chi Lai, Hung-Jen Chen, and Min Sun. 2025. Grounding-aware token pruning: Recovering from drastic performance drops in visual grounding caused by pruning. Preprint, arXiv:2506.21873. — [arXiv:2506.21873](https://arxiv.org/abs/2506.21873) find
  13. Mostafa Dehghani, Anurag Arnab, Lucas Beyer, Ashish Vaswani, and Yi Tay. 2022. The efficiency misnomer. Preprint, arXiv:2110.12894. find
  14. Mark Endo, Xiaohan Wang, and Serena Yeung-Levy. 2025. Feather the throttle: Revisiting visual token pruning for vision-language model acceleration. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 22826–22835. find
  15. Chaoyou Fu, Peixian Chen, Yunhang Shen, Yulei Qin, Mengdan Zhang, Xu Lin, Jinrui Yang, Xiawu Zheng, Ke Li, Xing Sun, and 1 others. 2025a. Mme: A comprehensive evaluation benchmark for multimodal large language models. In The Thirty-ninth Annual Conference on Neural Information Processing Systems Dat…
  16. Chaoyou Fu, Yuhan Dai, Yongdong Luo, Lei Li, Shuhuai Ren, Renrui Zhang, Zihan Wang, Chenyu Zhou, Yunhang Shen, Mengdan Zhang, and 1 others. 2025b. Video-mme: The first-ever comprehensive evaluation benchmark of multi-modal llms in video analysis. In Proceedings of the Computer Vision and Pattern Rec…
  17. Mingyu Fu, Wei Suo, Ji Ma, Lin Yuanbo Wu, Peng Wang, and Yanning Zhang. 2025c. Mitigating information loss under high pruning rates for efficient large vision language models. In Proceedings of the 33rd ACM International Conference on Multimedia, pages 4156–4165. find
  18. Tianyu Fu, Tengxuan Liu, Qinghao Han, Guohao Dai, Shengen Yan, Huazhong Yang, Xuefei Ning, and Yu Wang. 2025d. Framefusion: Combining similarity and importance for video token reduction on large vision language models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages… find
  19. Yash Goyal, Tejas Khot, Douglas Summers-Stay, Dhruv Batra, and Devi Parikh. 2016. Making the V in VQA matter: Elevating the role of image understanding in visual question answering. CoRR, abs/1612.00837.
  20. Tianrui Guan, Fuxiao Liu, Xiyang Wu, Ruiqi Xian, Zongxia Li, Xiaoyu Liu, Xijun Wang, Lichang Chen, Furong Huang, Yaser Yacoob, Dinesh Manocha, and Tianyi Zhou. 2024. Hallusionbench: An advanced diagnostic suite for entangled language hallucination and visual illusion in large vision-language models.…
  21. Yichen Guo, Hanze Li, Zonghao Zhang, Jinhao You, Kai Tang, and Xiande Huang. 2025. Star: Stage-wise attention-guided token reduction for efficient large vision-language models inference. Preprint, arXiv:2505.12359. — [arXiv:2505.12359](https://arxiv.org/abs/2505.12359) find
  22. Danna Gurari, Qing Li, Abigale J Stangl, Anhong Guo, Chi Lin, Kristen Grauman, Jiebo Luo, and Jeffrey P Bigham. 2018. Vizwiz grand challenge: Answering visual questions from blind people. In Proceedings of the IEEE conference on computer vision and pattern recognition, pages 3608–3617. find
  23. Yuhang Han, Xuyang Liu, Zihan Zhang, Pengxiang Ding, Junjie Chen, Donglin Wang, Honggang Chen, Qingsen Yan, and Siteng Huang. 2025. Filter, correlate, compress: Training-free token reduction for mllm acceleration. Preprint, arXiv:2411.17686. — [arXiv:2411.17686](https://arxiv.org/abs/2411.17686) find
  24. Sara Hooker. 2020. The hardware lottery. Preprint, arXiv:2009.06489. — [arXiv:2009.06489](https://arxiv.org/abs/2009.06489) find
  25. Lianyu Hu, Fanhua Shang, Wei Feng, and Liang Wan. 2025. Lightvlm: Acceleraing large multimodal models with pyramid token merging and kv cache compression. Preprint, arXiv:2509.00419. — [arXiv:2509.00419](https://arxiv.org/abs/2509.00419) find
  26. Lianyu Hu, Fanhua Shang, Liang Wan, and Wei Feng. 2024. illava: An image is worth fewer than 1/3 input tokens in large multimodal models. Preprint, arXiv:2412.06263. — [arXiv:2412.06263](https://arxiv.org/abs/2412.06263) find
  27. Kai Huang, Hao Zou, Ye Xi, BoChen Wang, Zhen Xie, and Liang Yu. 2024. Ivtp: Instruction-guided visual token pruning for large vision-language models. In Computer Vision – ECCV 2024: 18th European Conference, Milan, Italy, September 29–October 4, 2024, Proceedings, Part XVII, page 214–230, Berlin, He… find
  28. Drew A. Hudson and Christopher D. Manning. 2019. GQA: a new dataset for compositional question answering over real-world images. CoRR, abs/1902.09506.
  29. Lei Jiang, Weizhe Huang, Tongxuan Liu, Yuting Zeng, Jing Li, Lechao Cheng, and Xiaohua Xu. 2024. Fopru: Focal pruning for efficient large vision-language models. CoRR. find
  30. Lei Jiang, Zixun Zhang, Yuting Zeng, Chunzhao Xie, Tongxuan Liu, Zhen Li, Lechao Cheng, and Xiaohua Xu. 2025. DCP: Dual-cue pruning for efficient large vision-language models. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, pages 21202–21215, Suzhou, China.… find
  31. Aniruddha Kembhavi, Mike Salvato, Eric Kolve, Minjoon Seo, Hannaneh Hajishirzi, and Ali Farhadi. 2016. A diagram is worth a dozen images. In Computer Vision – ECCV 2016, pages 235–251, Cham. Springer International Publishing.
  32. Yulong Lei, Zishuo Wang, Jinglin Xu, and Yuxin Peng. 2025. Ctp2fic: From coarse-grained token pruning to fine-grained token clustering for lvlm inference acceleration (chinamm 2025). Available at SSRN 5545751. find
  33. Ao Li, Yuxiang Duan, Jinghui Zhang, Congbo Ma, Yutong Xie, Gustavo Carneiro, Mohammad Yaqub, and Hu Wang. 2025a. Transprune: Token transition pruning for efficient large vision-language model. Preprint, arXiv:2507.20630. — [arXiv:2507.20630](https://arxiv.org/abs/2507.20630) find
  34. Duo Li, Zuhao Yang, Xiaoqin Zhang, Ling Shao, and Shijian Lu. 2025b. Todre: Effective visual token pruning via token diversity and task relevance. Preprint, arXiv:2505.18757. — [arXiv:2505.18757](https://arxiv.org/abs/2505.18757) find
  35. Junnan Li, Dongxu Li, Silvio Savarese, and Steven Hoi. 2023a. Blip-2: Bootstrapping language-image pre-training with frozen image encoders and large language models. Preprint, arXiv:2301.12597.
  36. Junnan Li, Dongxu Li, Caiming Xiong, and Steven Hoi. 2022. Blip: Bootstrapping language-image pre-training for unified vision-language understanding and generation. In International conference on machine learning, pages 12888–12900. PMLR.
  37. Kaiyuan Li, Xiaoyue Chen, Chen Gao, Yong Li, and Xinlei Chen. 2025c. Balanced token pruning: Accelerating vision language models beyond local optimization. Preprint, arXiv:2505.22038. — [arXiv:2505.22038](https://arxiv.org/abs/2505.22038) find
  38. Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2023b. Visual instruction tuning. Advances in neural information processing systems, 36:34892– 34916.
  39. Kunchang Li, Yali Wang, Yinan He, Yizhuo Li, Yi Wang, Yi Liu, Zun Wang, Jilan Xu, Guo Chen, Ping Luo, and 1 others. 2024. Mvbench: A comprehensive multi-modal video understanding benchmark. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pages 22195–22206.
  40. Juntao Liu, Liqiang Niu, Wenchao Chen, Jie Zhou, and Fandong Meng. 2025. Laco: Efficient layer-wise compression of visual tokens for multimodal large language models. Preprint, arXiv:2507.02279. — [arXiv:2507.02279](https://arxiv.org/abs/2507.02279) find
  41. Yangfu Li, Hongjian Zhan, Tianyi Chen, Qi Liu, and Yue Lu. 2025d. Why 1 + 1 \< 1 in visual token pruning: Beyond naive integration via multi-objective balanced covering. Preprint, arXiv:2505.10118. — [arXiv:2505.10118](https://arxiv.org/abs/2505.10118) find
  42. Yanshu Li, Jianjiang Yang, Zhennan Shen, Ligong Han, Haoyan Xu, and Ruixiang Tang. 2025e. Catp: Contextually adaptive token pruning for efficient and enhanced multimodal in-context learning. Preprint, arXiv:2508.07871. — [arXiv:2508.07871](https://arxiv.org/abs/2508.07871) find
  43. Yifan Li, Yifan Du, Kun Zhou, Jinpeng Wang, Xin Zhao, and Ji-Rong Wen. 2023b. Evaluating object hallucination in large vision-language models. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 292–305, Singapore. Association for Computational Linguistic…
  44. Yinan Liang, Ziwei Wang, Xiuwei Xu, Jie Zhou, and Jiwen Lu. 2025a. Efficientllava: Generalizable auto-pruning for large vision-language models. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 9445–9454. find
  45. Shilong Liu, Hao Cheng, Haotian Liu, Hao Zhang, Feng Li, Tianhe Ren, Xueyan Zou, Jianwei Yang, Hang Su, Jun Zhu, and 1 others. 2024b. Llava-plus: Learning to use tools for creating multimodal agents. In European conference on computer vision, pages 126–142. Springer. find
  46. Ting Liu, Liangtao Shi, Richang Hong, Yue Hu, Quanjun Yin, and Linfeng Zhang. 2024c. Multi-stage vision token dropping: Towards efficient multimodal large language model. Preprint, arXiv:2411.10803. — [arXiv:2411.10803](https://arxiv.org/abs/2411.10803) find
  47. Yuan Liu, Haodong Duan, Yuanhan Zhang, Bo Li, Songyang Zhang, Wangbo Zhao, Yike Yuan, Jiaqi Wang, Conghui He, Ziwei Liu, Kai Chen, and Dahua Lin. 2024d. Mmbench: Is your multi-modal model an all-around player? In Computer Vision – ECCV 2024: 18th European Conference, Milan, Italy, September 29–Octob…
  48. Yuliang Liu, Zhang Li, Mingxin Huang, Biao Yang, Wenwen Yu, Chunyuan Li, Xucheng Yin, Cheng lin Liu, Lianwen Jin, and Xiang Bai. 2023c. Ocrbench: on the hidden mystery of ocr in large multimodal models. Science China Information Sciences, 67.
  49. Yuxuan Liang, Xu Li, Xiaolei Chen, Yi Zheng, Haotian Chen, Bin Li, and Xiangyang Xue. 2025b. Pyramid token pruning for high-resolution large vision-language models via region, token, and instruction-guided importance. Preprint, arXiv:2509.15704. — [arXiv:2509.15704](https://arxiv.org/abs/2509.15704) find
  50. Pan Lu, Swaroop Mishra, Tanglin Xia, Liang Qiu, Kai-Wei Chang, Song-Chun Zhu, Oyvind Tafjord, Peter Clark, and Ashwin Kalyan. 2022. Learn to explain: Multimodal reasoning via thought chains for science question answering. Advances in Neural Information Processing Systems, 35:2507–2521. find
  51. Chaofan Lin, Jiaming Tang, Shuo Yang, Hanshuo Wang, Tian Tang, Boyu Tian, Ion Stoica, Song Han, and Mingyu Gao. 2025. Twilight: Adaptive attention sparsity with hierarchical top-p pruning. Preprint, arXiv:2502.02770. find
  52. Bozhi Luan, Wengang Zhou, Hao Feng, Zhe Wang, Xiaosong Li, and Houqiang Li. 2025. Multi-cue adaptive visual token pruning for large vision-language models. Preprint, arXiv:2503.08019. — [arXiv:2503.08019](https://arxiv.org/abs/2503.08019) find
  53. Tsung-Yi Lin, Michael Maire, Serge Belongie, Lubomir Bourdev, Ross Girshick, James Hays, Pietro Perona, Deva Ramanan, C. Lawrence Zitnick, and Piotr Dollár. 2015. Microsoft coco: Common objects in context. Preprint, arXiv:1405.0312.
  54. Haotian Liu, Chunyuan Li, Yuheng Li, and Yong Jae Lee. 2024a. Improved baselines with visual instruction tuning. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 26296–26306.
  55. Haotian Liu, Chunyuan Li, Qingyang Wu, and Yong Jae Lee. 2023a. Visual instruction tuning. Advances in neural information processing systems, 36:34892– 34916.
  56. Deng Luo, Dongyang Zhang, Qiuhao Xie, Cencen Liu, Qiang Dong, and Xiurui Xie. 2025. Rethinking attention cues: Multi-factor guided token pruning for efficient vision-language understanding. Available at SSRN 5615684. find
  57. Ji Ma, Wei Suo, Peng Wang, and Yanning Zhang. 2025a. Short-lvlm: Compressing and accelerating large vision-language models by pruning redundant layers. In Proceedings of the 33rd ACM International Conference on Multimedia, pages 3575–3584. find
  58. Junpeng Ma, Qizhe Zhang, Ming Lu, Zhibin Wang, Qiang Zhou, Jun Song, and Shanghang Zhang. 2025b. Mmg-vid: Maximizing marginal gains at segment-level and token-level for efficient video llms. Preprint, arXiv:2508.21044. — [arXiv:2508.21044](https://arxiv.org/abs/2508.21044) find
  59. Ahmed Masry, Xuan Long Do, Jia Qing Tan, Shafiq Joty, and Enamul Hoque. 2022. Chartqa: A benchmark for question answering about charts with visual and logical reasoning. In Findings of the association for computational linguistics: ACL 2022, pages 2263– 2279.
  60. Wei Suo, Ji Ma, Mengyang Sun, Lin Yuanbo Wu, Peng Wang, and Yanning Zhang. 2025. Pruning all-rounder: Rethinking and improving inference efficiency for large vision language models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 20247–20256. find
  61. Minesh Mathew, Dimosthenis Karatzas, R Manmatha, and CV Jawahar. 2020. Docvqa: A dataset for vqa on document images. corr abs/2007.00398 (2020). arXiv preprint arXiv:2007.00398.
  62. Peng Wang, Shuai Bai, Sinan Tan, Shijie Wang, Zhihao Fan, Jinze Bai, Keqin Chen, Xuejing Liu, Jialin Wang, Wenbin Ge, and 1 others. 2024. Qwen2-vl: Enhancing vision-language model’s perception of the world at any resolution. arXiv preprint arXiv:2409.12191.
  63. Yu Meng, Kaiyuan Li, Chenran Huang, Chen Gao, Xinlei Chen, Yong Li, and Xiaoping Zhang. 2025. Plphp: Per-layer per-head vision token pruning for efficient large vision-language models. Preprint, arXiv:2502.14504. — [arXiv:2502.14504](https://arxiv.org/abs/2502.14504) find
  64. A. Mishra, K. Alahari, and C. V. Jawahar. 2012. Scene text recognition using higher order language priors. In BMVC. find
  65. Ruiguang Pei, Weiqing Sun, Zhihui Fu, and Jun Wang. 2025. Greedyprune: Retenting critical visual token set for large vision language models. Preprint, arXiv:2506.13166. — [arXiv:2506.13166](https://arxiv.org/abs/2506.13166) find
  66. Birgit Pfitzmann, Christoph Auer, Michele Dolfi, Ahmed S Nassar, and Peter W J Staar. 2022. Doclaynet: A large human-annotated dataset for document-layout segmentation. page 3743–3751. find
  67. Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, Gretchen Krueger, and Ilya Sutskever. 2021. Learning transferable visual models from natural language supervision. Preprint, arXiv:2103.00020.
  68. Yuzhang Shang, Mu Cai, Bingxin Xu, Yong Jae Lee, and Yan Yan. 2025. Llava-prumerge: Adaptive token reduction for efficient large multimodal models. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 22857–22867. find
  69. Kele Shao, Keda Tao, Kejia Zhang, Sicheng Feng, Mu Cai, Yuzhang Shang, Haoxuan You, Can Qin, Yang Sui, and Huan Wang. 2025. When tokens talk too much: A survey of multimodal long-context token compression across images, videos, and audios. arXiv preprint arXiv:2507.20198. — [arXiv:2507.20198](https://arxiv.org/abs/2507.20198) find
  70. Amanpreet Singh, Vivek Natarajan, Meet Shah, Yu Jiang, Xinlei Chen, Dhruv Batra, Devi Parikh, and Marcus Rohrbach. 2019. Towards vqa models that can read. In Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pages 8317–8326. find
  71. Yi-Lin Sung, Jaehong Yoon, and Mohit Bansal. 2024. Ecoflap: Efficient coarse-to-fine layer-wise pruning for vision-language models. Preprint, arXiv:2310.02998. find
  72. Tiannan Wang, Wangchunshu Zhou, Yan Zeng, and Xinsong Zhang. 2023. Efficientvlm: Fast and accurate vision-language models via knowledge distillation and modal-adaptive pruning. In Findings of the association for computational linguistics: ACL 2023, pages 13899–13913. find
  73. Zichen Wen, Yifeng Gao, Shaobo Wang, Junyuan Zhang, Qintong Zhang, Weijia Li, Conghui He, and Linfeng Zhang. 2025. Stop looking for important tokens in multimodal language models: Duplication matters more. Preprint, arXiv:2502.11494.
  74. Haoning Wu, Dongxu Li, Bei Chen, and Junnan Li. 2024. Longvideobench: A benchmark for long-context interleaved video-language understanding. Advances in Neural Information Processing Systems, 37:28828–28857.
  75. Zimeng Wu, Jiaxin Chen, and Yunhong Wang. 2025. Unified knowledge maintenance pruning and progressive recovery with weight recalling for large vision-language models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 8550–8558. find
  76. Long Xing, Qidong Huang, Xiaoyi Dong, Jiajie Lu, Pan Zhang, Yuhang Zang, Yuhang Cao, Conghui He, Jiaqi Wang, Feng Wu, and Dahua Lin. 2025. Pyramiddrop: Accelerating your large vision-language models via pyramid visual redundancy reduction. Preprint, arXiv:2410.17247. — [arXiv:2410.17247](https://arxiv.org/abs/2410.17247) find
  77. Rui Xu, Yunke Wang, Yong Luo, and Bo Du. 2025. Rethinking visual token reduction in lvlms under cross-modal misalignment. Preprint, arXiv:2506.22283. — [arXiv:2506.22283](https://arxiv.org/abs/2506.22283) find
  78. Ran Yan, Youhe Jiang, Zhuoming Chen, Haohui Mai, Beidi Chen, and Binhang Yuan. 2025. Fsa: An alternative efficient implementation of native sparse attention kernel. Preprint, arXiv:2508.18224. find
  79. Longrong Yang, Dong Shen, Chaoxiang Cai, Kaibing Chen, Fan Yang, Tingting Gao, Di Zhang, and Xi Li. 2025a. Libra-merging: Importance-redundancy and pruning-merging trade-off for acceleration plug-in in large vision-language model. In 2025 IEEE/CVF Conference on Computer Vision and Pattern Recognitio… find
  80. Senqiao Yang, Yukang Chen, Zhuotao Tian, Chengyao Wang, Jingyao Li, Bei Yu, and Jiaya Jia. 2025b. Visionzip: Longer is better but not necessary in vision language models. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 19792–19802. find
  81. Junyang Zhang, Mu Yuan, Ruiguang Zhong, Puhan Luo, Huiyou Zhan, Ningkang Zhang, Chengchen Hu, and Xiang-Yang Li. 2025b. A-vl: Adaptive attention for large vision-language models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pages 22461–22469. find
  82. Huanjin Yao, Ruifei Zhang, Jiaxing Huang, Jingyi Zhang, Yibo Wang, Bo Fang, Ruolin Zhu, Yongcheng Jing, Shunyu Liu, Guanbin Li, and Dacheng Tao. 2025. A survey on agentic multimodal large language models. Preprint, arXiv:2510.10991. — [arXiv:2510.10991](https://arxiv.org/abs/2510.10991) find
  83. Qizhe Zhang, Aosong Cheng, Ming Lu, Renrui Zhang, Zhiyong Zhuo, Jiajun Cao, Shaobo Guo, Qi She, and Shanghang Zhang. 2025c. Beyond text-visual attention: Exploiting visual cues for effective token pruning in vlms. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 2085… find
  84. Weihao Ye, Qiong Wu, Wenhao Lin, and Yiyi Zhou. 2025a. Fit and prune: fast and training-free visual token pruning for multi-modal large language models. In Proceedings of the Thirty-Ninth AAAI Conference on Artificial Intelligence and Thirty-Seventh Conference on Innovative Applications of Artificia… find
  85. Xubing Ye, Yukang Gan, Yixiao Ge, Xiao-Ping Zhang, and Yansong Tang. 2025b. Atp-llava: Adaptive token pruning for large vision language models. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 24972–24982. find
  86. Zihao Ye, Lequn Chen, Ruihang Lai, Wuwei Lin, Yi-neng Zhang, Stephanie Wang, Tianqi Chen, Baris Kasikci, Vinod Grover, Arvind Krishnamurthy, and Luis Ceze. 2025c. Flashinfer: Efficient and customizable attention engine for llm inference serving. Preprint, arXiv:2501.01005.
  87. Peter Young, Alice Lai, Micah Hodosh, and Julia Hockenmaier. 2014. From image descriptions to visual denotations: New similarity metrics for semantic inference over event descriptions. Transactions of the Association for Computational Linguistics, 2:67–78. find
  88. Weihao Yu, Zhengyuan Yang, Linjie Li, Jianfeng Wang, Kevin Lin, Zicheng Liu, Xinchao Wang, and Lijuan Wang. 2023. Mm-vet: Evaluating large multimodal models for integrated capabilities. arXiv preprint arXiv:2308.02490.
  89. Quan-Sheng Zeng, Yunheng Li, Qilong Wang, Peng-Tao Jiang, Zuxuan Wu, Ming-Ming Cheng, and Qibin Hou. 2025. A glimpse to compress: Dynamic visual token pruning for large vision-language models. Preprint, arXiv:2508.01548. — [arXiv:2508.01548](https://arxiv.org/abs/2508.01548) find
  90. Xiaohua Zhai, Basil Mustafa, Alexander Kolesnikov, and Lucas Beyer. 2023. Sigmoid loss for language image pre-training. Preprint, arXiv:2303.15343.
  91. Ce Zhang, Kaixin Ma, Tianqing Fang, Wenhao Yu, Hongming Zhang, Zhisong Zhang, Yaqi Xie, Katia Sycara, Haitao Mi, and Dong Yu. 2025a. Vscan: Rethinking visual token reduction for efficient large vision-language models. Preprint, arXiv:2505.22654. — [arXiv:2505.22654](https://arxiv.org/abs/2505.22654) find
  92. Qizhe Zhang, Mengzhen Liu, Lichen Li, Ming Lu, Yuan Zhang, Junwen Pan, Qi She, and Shanghang Zhang. 2025d. Beyond attention or similarity: Maximizing conditional diversity for token pruning in mllms. Preprint, arXiv:2506.10967. find
  93. Yuan Zhang, Chun-Kai Fan, Junpeng Ma, Wenzhao Zheng, Tao Huang, Kuan Cheng, Denis Gudovskiy, Tomoyuki Okuno, Yohei Nakata, Kurt Keutzer, and Shanghang Zhang. 2025e. Sparsevlm: Visual token sparsification for efficient vision-language model inference. Preprint, arXiv:2410.04417. find
  94. Zhenyu Zhang, Ying Sheng, Tianyi Zhou, Tianlong Chen, Lianmin Zheng, Ruisi Cai, Zhao Song, Yuandong Tian, Christopher Ré, Clark Barrett, and 1 others. 2023. H2o: Heavy-hitter oracle for efficient generative inference of large language models. Advances in Neural Information Processing Systems, 36:346… find
  95. Yiwu Zhong, Zhuoming Liu, Yin Li, and Liwei Wang. 2025. Aim: Adaptive inference of multi-modal llms via token merging and pruning. In Proceedings of the IEEE/CVF International Conference on Computer Vision, pages 20180–20192. find
  96. Junjie Zhou, Yan Shu, Bo Zhao, Boya Wu, Zhengyang Liang, Shitao Xiao, Minghao Qin, Xi Yang, Yongping Xiong, Bo Zhang, and 1 others. 2025. Mlvu: Benchmarking multi-task long video understanding. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 13691–13701.
  97. Xunyu Zhu, Jian Li, Yong Liu, Can Ma, and Weiping Wang. 2024. A survey on model compression for large language models. Transactions of the Association for Computational Linguistics, 12:1556–1577.
  98. Xianwei Zhuang, Zhihong Zhu, Yuxin Xie, Liming Liang, and Yuexian Zou. 2025. Vasparse: Towards efficient visual hallucination mitigation via visual-aware token sparsification. In Proceedings of the Computer Vision and Pattern Recognition Conference, pages 4189–4199. find
  99. Xin Zou, Di Lu, Yizhou Wang, Yibo Yan, Yuanhuiyi Lyu, Xu Zheng, Linfeng Zhang, and Xuming Hu. 2025. Don’t just chase "highlighted tokens" in mllms: Revisiting visual holistic context retention. Preprint, arXiv:2510.02912. — [arXiv:2510.02912](https://arxiv.org/abs/2510.02912) find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/yuan-2026-reducing-token-redundancy-lvlms-systematic/