# Visualizing and Understanding Patch Interactions in Vision Transformer

arXiv: https://arxiv.org/abs/2203.05922 · HTML: https://mlsanity.com/paper/2203.05922/

## Key facts

- Cited by 26 papers in the map (data through 14 Oct 2026).
- Builds on 37 of its 56 references that are in the map.
- First posted on arXiv 11 Mar 2022.

## Builds on (most cited first, 30 of 37)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Deformable DETR: Deformable Transformers for End-to-End Object Detection](https://mlsanity.com/paper/2010.04159/) (2020)
- [Axiomatic Attribution for Deep Networks](https://mlsanity.com/paper/1703.01365/) (2017)
- [SegFormer: Simple and Efficient Design for Semantic Segmentation with Transformers](https://mlsanity.com/paper/2105.15203/) (2021)
- [BEiT: BERT Pre-Training of Image Transformers](https://mlsanity.com/paper/2106.08254/) (2021)
- [Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions](https://mlsanity.com/paper/2102.12122/) (2021)
- [Rethinking Semantic Segmentation from a Sequence-to-Sequence Perspective with Transformers](https://mlsanity.com/paper/2012.15840/) (2020)
- [SmoothGrad: removing noise by adding noise](https://mlsanity.com/paper/1706.03825/) (2017)
- [What Does BERT Look At? An Analysis of BERT's Attention](https://mlsanity.com/paper/1906.04341/) (2019)
- [DynamicViT: Efficient Vision Transformers with Dynamic Token Sparsification](https://mlsanity.com/paper/2106.02034/) (2021)
- [Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Pruned](https://mlsanity.com/paper/1905.09418/) (2019)
- [Are Sixteen Heads Really Better than One?](https://mlsanity.com/paper/1905.10650/) (2019)
- [A Primer in BERTology: What we know about how BERT works](https://mlsanity.com/paper/2002.12327/) (2020)
- [Do Vision Transformers See Like Convolutional Neural Networks?](https://mlsanity.com/paper/2108.08810/) (2021)
- [CSWin Transformer: A General Vision Transformer Backbone with Cross-Shaped Windows](https://mlsanity.com/paper/2107.00652/) (2021)
- [Intriguing Properties of Vision Transformers](https://mlsanity.com/paper/2105.10497/) (2021)
- [Is Attention Interpretable?](https://mlsanity.com/paper/1906.03731/) (2019)
- [Not Just a Black Box: Learning Important Features Through Propagating Activation Differences](https://mlsanity.com/paper/1605.01713/) (2016)
- [Explaining NonLinear Classification Decisions with Deep Taylor Decomposition](https://mlsanity.com/paper/1512.02479/) (2015)
- [LocalViT: Analyzing Locality in Vision Transformers](https://mlsanity.com/paper/2104.05707/) (2021)
- [Full-Gradient Representation for Neural Network Visualization](https://mlsanity.com/paper/1905.00780/) (2019)
- [Rethinking Transformer-based Set Prediction for Object Detection](https://mlsanity.com/paper/2011.10881/) (2020)
- [Salient Object Detection via Integrity Learning](https://mlsanity.com/paper/2101.07663/) (2021)
- [Character-Aware Neural Language Models](https://mlsanity.com/paper/1508.06615/) (2015)
- [TransFG: A Transformer Architecture for Fine-grained Recognition](https://mlsanity.com/paper/2103.07976/) (2021)

## Built on by (most cited first, 26 of 26)

- [Explainable and Interpretable Multimodal Large Language Models: A Comprehensive Survey](https://mlsanity.com/paper/2412.02104/) (2024)
- [AttentionViz: A Global View of Transformer Attention](https://mlsanity.com/paper/2305.03210/) (2023)
- [Hierarchical Local-Global Transformer for Temporal Sentence Grounding](https://mlsanity.com/paper/2208.14882/) (2022)
- [QSAN: A Near-term Achievable Quantum Self-Attention Network](https://mlsanity.com/paper/2207.07563/) (2022)
- [Pay Attention and Move Better: Harnessing Attention for Interactive Motion Generation and Training-free Editing](https://mlsanity.com/paper/2410.18977/) (2024)
- [Explainability of Vision Transformers: A Comprehensive Review and New Perspectives](https://mlsanity.com/paper/2311.06786/) (2023)
- [Application of Multimodal Fusion Deep Learning Model in Disease Recognition](https://mlsanity.com/paper/2406.18546/) (2024)
- [Transformer Explainer: Learning LLM Transformers with Interactive Visual Explanation and Experimentation](https://mlsanity.com/paper/2408.04619/) (2024)
- [Understanding Multi-View Transformers](https://mlsanity.com/paper/2510.24907/) (2025)
- [Recasting Generic Pretrained Vision Transformers As Object-Centric Scene Encoders For Manipulation Policies](https://mlsanity.com/paper/2405.15916/) (2024)
- [FViT: A Focal Vision Transformer with Gabor Filter](https://mlsanity.com/paper/2402.11303/) (2024)
- [Demystify Self-Attention in Vision Transformers from a Semantic Perspective: Analysis and Application](https://mlsanity.com/paper/2211.08543/) (2022)
- [Improved Ear Verification with Vision Transformers and Overlapping Patches](https://mlsanity.com/paper/2503.23275/) (2025)
- [Knowledge Transfer from Interaction Learning](https://mlsanity.com/paper/2509.18733/) (2025)
- [ForensicConcept: Transferable Forensic Concepts for AIGI Detection](https://mlsanity.com/paper/2606.07034/) (2026)
- [Universal Deoxidation of Semiconductor Substrates Assisted by Machine-Learning and Real-Time-Feedback-Control](https://mlsanity.com/paper/2312.01662/) (2023)
- [VisMoDAl: Visual Analytics for Evaluating and Improving Corruption Robustness of Vision-Language Models](https://mlsanity.com/paper/2509.14571/) (2025)
- [An Evaluation of Hybrid Annotation Workflows on High-Ambiguity Spatiotemporal Video Footage](https://mlsanity.com/paper/2510.21798/) (2025)
- [Evading Visual Aphasia: Contrastive Adaptive Semantic Token Pruning for Vision-Language Models](https://mlsanity.com/paper/2605.09429/) (2026)
- [Model-Contrastive Federated Domain Adaptation](https://mlsanity.com/paper/2305.10432/) (2023)
- [ClickAttention: Click Region Similarity Guided Interactive Segmentation](https://mlsanity.com/paper/2408.06021/) (2024)
- [Learning Object Focused Attention](https://mlsanity.com/paper/2504.08166/) (2025)
- [PerFormer: A Permutation Based Vision Transformer for Remaining Useful Life Prediction](https://mlsanity.com/paper/2506.00259/) (2025)
- [MoDA: Modulation Adapter for Fine-Grained Visual Grounding in Instructional MLLMs](https://mlsanity.com/paper/2506.01850/) (2025)
- [PGKET: A Photonic Gaussian Kernel Enhanced Transformer](https://mlsanity.com/paper/2507.19041/) (2025)
- [ViT-Explainer: An Interactive Walkthrough of the Vision Transformer Pipeline](https://mlsanity.com/paper/2604.02182/) (2026)

## Abstract (arXiv)

Vision Transformer (ViT) has become a leading tool in various computer vision tasks, owing to its unique self-attention mechanism that learns visual representations explicitly through cross-patch information interactions. Despite having good success, the literature seldom explores the explainability of vision transformer, and there is no clear picture of how the attention mechanism with respect to the correlation across comprehensive patches will impact the performance and what is the further potential. In this work, we propose a novel explainable visualization approach to analyze and interpret the crucial attention interactions among patches for vision transformer. Specifically, we first introduce a quantification indicator to measure the impact of patch interaction and verify such quantification on attention window design and indiscriminative patches removal. Then, we exploit the effective responsive field of each patch in ViT and devise a window-free transformer architecture accordingly. Extensive experiments on ImageNet demonstrate that the exquisitely designed quantitative method is shown able to facilitate ViT model learning, leading the top-1 accuracy by 4.28% at most. Moreover, the results on downstream fine-grained recognition tasks further validate the generalization of our proposal.