# Beyond the Doors of Perception: Vision Transformers Represent Relations Between Objects

arXiv: https://arxiv.org/abs/2406.15955 · HTML: https://mlsanity.com/paper/2406.15955/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- Builds on 46 of its 76 references that are in the map.
- First posted on arXiv 22 Jun 2024.

## Builds on (most cited first, 30 of 46)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [ImageNet Large Scale Visual Recognition Challenge](https://mlsanity.com/paper/1409.0575/) (2014)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization](https://mlsanity.com/paper/1610.02391/) (2016)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps](https://mlsanity.com/paper/1312.6034/) (2013)
- [Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small](https://mlsanity.com/paper/2211.00593/) (2022)
- [Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations](https://mlsanity.com/paper/1811.12359/) (2018)
- [Vision Transformers Need Registers](https://mlsanity.com/paper/2309.16588/) (2023)
- [Progress measures for grokking via mechanistic interpretability](https://mlsanity.com/paper/2301.05217/) (2023)
- [A-OKVQA: A Benchmark for Visual Question Answering using World Knowledge](https://mlsanity.com/paper/2206.01718/) (2022)
- [Do Vision Transformers See Like Convolutional Neural Networks?](https://mlsanity.com/paper/2108.08810/) (2021)
- [When and why vision-language models behave like bags-of-words, and what to do about it?](https://mlsanity.com/paper/2210.01936/) (2022)
- [Transformer Feed-Forward Layers Build Predictions by Promoting Concepts in the Vocabulary Space](https://mlsanity.com/paper/2203.14680/) (2022)
- [Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality](https://mlsanity.com/paper/2204.03162/) (2022)
- [Emergent World Representations: Exploring a Sequence Model Trained on a Synthetic Task](https://mlsanity.com/paper/2210.13382/) (2022)
- [Emergent Linear Representations in World Models of Self-Supervised Sequence Models](https://mlsanity.com/paper/2309.00941/) (2023)
- [Towards a Definition of Disentangled Representations](https://mlsanity.com/paper/1812.02230/) (2018)
- [Visual Spatial Reasoning](https://mlsanity.com/paper/2205.00363/) (2022)
- [Multi-Grained Vision Language Pre-Training: Aligning Texts with Visual Concepts](https://mlsanity.com/paper/2111.08276/) (2021)
- [Finding Alignments Between Interpretable Causal Variables and Distributed Neural Representations](https://mlsanity.com/paper/2303.02536/) (2023)
- [How does GPT-2 compute greater-than?: Interpreting mathematical abilities in a pre-trained language model](https://mlsanity.com/paper/2305.00586/) (2023)
- [CRAFT: Concept Recursive Activation FacTorization for Explainability](https://mlsanity.com/paper/2211.10154/) (2022)
- [How much do language models copy from their training data? Evaluating linguistic novelty in text generation using RAVEN](https://mlsanity.com/paper/2111.09509/) (2021)
- [Circuit Component Reuse Across Tasks in Transformer Language Models](https://mlsanity.com/paper/2310.08744/) (2023)
- [How do Language Models Bind Entities in Context?](https://mlsanity.com/paper/2310.17191/) (2023)

## Built on by (most cited first, 13 of 13)

- [When Models Manipulate Manifolds: The Geometry of a Counting Task](https://mlsanity.com/paper/2601.04480/) (2026)
- [How Do Vision-Language Models Process Conflicting Information Across Modalities?](https://mlsanity.com/paper/2507.01790/) (2025)
- [Visual symbolic mechanisms: Emergent symbol processing in vision language models](https://mlsanity.com/paper/2506.15871/) (2025)
- [Whither symbols in the era of advanced neural networks?](https://mlsanity.com/paper/2508.05776/) (2025)
- [LinkedOut: Linking World Knowledge Representation Out of Video LLM for Next-Generation Video Recommendation](https://mlsanity.com/paper/2512.16891/) (2025)
- [Signatures of human-like processing in Transformer forward passes](https://mlsanity.com/paper/2504.14107/) (2025)
- [Self-Distillation of Hidden Layers for Self-Supervised Representation Learning](https://mlsanity.com/paper/2603.15553/) (2026)
- [Causality neq Decodability, and Vice Versa: Lessons from Interpreting Counting ViTs](https://mlsanity.com/paper/2510.09794/) (2025)
- [I Walk the Line: Examining the Role of Gestalt Continuity in Object Binding for Vision Transformers](https://mlsanity.com/paper/2604.09942/) (2026)
- [BlueGlass: A Framework for Composite AI Safety](https://mlsanity.com/paper/2507.10106/) (2025)
- [Linguistic Context Recodes Visual Representations in Vision-Language Models](https://mlsanity.com/paper/2608.00035/) (2026)
- [Intuition vectors](https://mlsanity.com/paper/2609.32619/) (2026)
- [Metonymic Circuits for Abstract Concept Grounding in Vision Transformers](https://mlsanity.com/paper/2610.06928/) (2026)

## Abstract (arXiv)

Though vision transformers (ViTs) have achieved state-of-the-art performance in a variety of settings, they exhibit surprising failures when performing tasks involving visual relations. This begs the question: how do ViTs attempt to perform tasks that require computing visual relations between objects? Prior efforts to interpret ViTs tend to focus on characterizing relevant low-level visual features. In contrast, we adopt methods from mechanistic interpretability to study the higher-level visual algorithms that ViTs use to perform abstract visual reasoning. We present a case study of a fundamental, yet surprisingly difficult, relational reasoning task: judging whether two visual entities are the same or different. We find that pretrained ViTs fine-tuned on this task often exhibit two qualitatively different stages of processing despite having no obvious inductive biases to do so: 1) a perceptual stage wherein local object features are extracted and stored in a disentangled representation, and 2) a relational stage wherein object representations are compared. In the second stage, we find evidence that ViTs can learn to represent somewhat abstract visual relations, a capability that has long been considered out of reach for artificial neural networks. Finally, we demonstrate that failures at either stage can prevent a model from learning a generalizable solution to our fairly simple tasks. By understanding ViTs in terms of discrete processing stages, one can more precisely diagnose and rectify shortcomings of existing and future models.