mlsanityData to 7 Oct 2026

MI-Pruner: Crossmodal Mutual Information-guided Token Pruner for Efficient MLLMs

arXiv 2604.03072 ↗First posted 3 Apr 2026cs.CVcited by 2

On the map: Multimodal LLMs / vision-language models → Visual token pruning and efficiency

Abstract (arXiv)

For multimodal large language models (MLLMs), visual information is relatively sparse compared with text. As a result, research on visual pruning emerges for efficient inference. Current approaches typically measure token importance based on the attention scores in the visual encoder or in the LLM decoder, then select visual tokens with high attention scores while pruning others. In this paper, we pursue a different and more surgical approach. Instead of relying on mechanism-specific signals, we directly compute Mutual Information (MI) between visual and textual features themselves, prior to their interaction. This allows us to explicitly measure crossmodal dependency at the feature levels. Our MI-Pruner is simple, efficient and non-intrusive, requiring no access to internal attention maps or architectural modifications. Experimental results demonstrate that our approach outperforms previous attention-based pruning methods with minimal latency.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Placed in Visual token pruning and efficiency, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 55 of its 76 references that are in the map. First posted on arXiv 3 Apr 2026.

Builds on

55 of 76 references.

Closest ancestors · 19

Same direction, newest first

  1. 2025FlashVLM: Text-Guided Visual Token Selection for Large Multimodal Models23 Dec
  2. 2025AutoPrune: Each Complexity Deserves a Pruning Policy28 Sep
  3. 2025MMTok: Multimodal Coverage Maximization for Efficient Inference of VLMs25 Aug
  4. 2025Beyond Attention or Similarity: Maximizing Conditional Diversity for Token Pruning in MLLMs12 Jun
  5. 2025Lifting the Veil on Visual Information Flow in MLLMs: Unlocking Pathways to Faster Inference17 Mar
  6. 2025Similarity-Aware Token Pruning: Your VLM but Faster14 Mar
  7. 2025See What You Are Told: Visual Attention Sink in Large Multimodal Models5 Mar
  8. 2025DivPrune: Diversity-based Visual Token Pruning for Large Multimodal Models4 Mar
  9. 2025Token Pruning in Multimodal Large Language Models: Are We Solving the Right Problem?17 Feb
  10. 2025Stop Looking for Important Tokens in Multimodal Language Models: Duplication Matters More17 Feb
  11. 2024VisionZip: Longer is Better but Not Necessary in Vision Language Models5 Dec
  12. 2024Beyond Text-Visual Attention: Exploiting Visual Cues for Effective Token Pruning in VLMs2 Dec
7 more
  1. 2024Accelerating Multimodal Large Language Models by Searching Optimal Vision Token Reduction30 Nov
  2. 2024ATP-LLaVA: Adaptive Token Pruning for Large Vision Language Models30 Nov
  3. 2024SparseVLM: Visual Token Sparsification for Efficient Vision-Language Model Inference6 Oct
  4. 2024Boosting Multimodal Large Language Models with Visual Tokens Withdrawal for Rapid Inference9 May
  5. 2024MULTIFLOW: Shifting Towards Task-Agnostic Vision-Language Pruning8 Apr
  6. 2024An Image is Worth 1/2 Tokens After Layer 2: Plug-and-Play Inference Acceleration for Large Vision-Language Models11 Mar
  7. 2022EfficientVLM: Fast and Accurate Vision-Language Models via Knowledge Distillation and Modal-adaptive Pruning14 Oct

Other references · 36

  1. 2014Microsoft COCO: Common Objects in Contextcited by 17,498
  2. 2023BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Modelscited by 6,069
  3. 2023Improved Baselines with Visual Instruction Tuningcited by 4,362
  4. 2024Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolutioncited by 4,151
  5. 2022FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awarenesscited by 3,611
  6. 2024LLaVA-OneVision: Easy Visual Task Transfercited by 2,669
  7. 2023FlashAttention-2: Faster Attention with Better Parallelism and Work Partitioningcited by 2,551
  8. 2023InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Taskscited by 2,534
  9. 2025Qwen3-VL Technical Reportcited by 2,534
  10. 2023Efficient Streaming Language Models with Attention Sinkscited by 2,115
  11. 2022Learn to Explain: Multimodal Reasoning via Thought Chains for Science Question Answeringcited by 1,881
  12. 2019Towards VQA Models That Can Readcited by 1,696
24 more
  1. 2023Evaluating Object Hallucination in Large Vision-Language Modelscited by 1,696
  2. 2023MME: A Comprehensive Evaluation Benchmark for Multimodal Large Language Modelscited by 1,624
  3. 2024Expanding Performance Boundaries of Open-Source Multimodal Models with Model, Data, and Test-Time Scalingcited by 1,573
  4. 2023Video-LLaVA: Learning United Visual Representation by Alignment Before Projectioncited by 1,385
  5. 2023MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilitiescited by 1,221
  6. 2019Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Prunedcited by 844
  7. 2023Vision Transformers Need Registerscited by 844
  8. 2023Mitigating Object Hallucinations in Large Vision-Language Models through Visual Contrastive Decodingcited by 671
  9. 2024Hallucination of Multimodal Large Language Models: A Surveycited by 380
  10. 2017TGIF-QA: Toward Spatio-Temporal Reasoning in Visual Question Answeringcited by 378
  11. 2024What matters when building vision-language models?cited by 321
  12. 2023Detecting and Preventing Hallucinations in Large Vision Language Modelscited by 254
  13. 2020SpAtten: Efficient Sparse Attention Architecture with Cascade Token and Head Pruningcited by 228
  14. 2024Multi-Modal Hallucination Control by Visual Information Groundingcited by 196
  15. 2016TGIF: A New Dataset and Benchmark on Animated GIF Descriptioncited by 124
  16. 2025Your Large Vision-Language Model Only Needs A Few Attention Heads For Visual Groundingcited by 78
  17. 2024Cross-modal Information Flow in Multimodal Large Language Modelscited by 65
  18. 2024MADTP: Multimodal Alignment-Guided Dynamic Token Pruning for Accelerating Vision-Language Transformercited by 53
  19. 2022Mutual Information Divergence: A Unified Metric for Multimodal Generative Modelscited by 35
  20. 2021Layer-wise Model Pruning based on Mutual Informationcited by 19
  21. 2025Hallucinatory Image Tokens: A Training-free EAZY Approach on Detecting and Mitigating Object Hallucinations in LVLMscited by 14
  22. 2025VisiPruner: Decoding Discontinuous Cross-Modal Dynamics for Efficient Multimodal LLMscited by 14
  23. 2025Vision Transformers with Self-Distilled Registerscited by 12
  24. 2025Not All Tokens and Heads Are Equally Important: Dual-Level Attention Intervention for Hallucination Mitigationcited by 7

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026EchoPrune: Interpreting Redundancy as Temporal Echoes for Efficient VideoLLMscited by 1
  2. 2026C-PTQ: Fisher-weighted Channel-wise Sensitivity for Post-training Quantization of MLLMscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026SinkPruner: Sink-Free Visual Token Pruning for Multimodal Large Language Models
  2. 2026IDPruner: Harmonizing Importance and Diversity in Visual Token Pruning for MLLMs
  3. 2024Fit and Prune: Fast and Training-free Visual Token Pruning for Multi-modal Large Language Models
  4. 2025Don't Just Chase "Highlighted Tokens" in MLLMs: Revisiting Visual Holistic Context Retention
  5. 2026EntropyPrune: Matrix Entropy Guided Visual Token Pruning for Multimodal Large Language Models
  6. 2025VFlowOpt: A Token Pruning Framework for LMMs with Visual Information Flow-Guided Optimization
  7. 2024ST3: Accelerating Multimodal Large Language Model by Spatial-Temporal Visual Token Trimming
  8. 2026SepPrune:A Separator-based Pruning Framework for Efficient Multimodal Large Language Models