# Does CLIP Bind Concepts? Probing Compositionality in Large Image Models

arXiv: https://arxiv.org/abs/2212.10537 · HTML: https://mlsanity.com/paper/2212.10537/

## Key facts

- Cited by 102 papers in the map (data through 14 Oct 2026).
- Builds on 42 of its 85 references that are in the map.
- First posted on arXiv 20 Dec 2022.

## Builds on (most cited first, 30 of 42)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [BLIP: Bootstrapping Language-Image Pre-training for Unified Vision-Language Understanding and Generation](https://mlsanity.com/paper/2201.12086/) (2022)
- [Distributed Representations of Words and Phrases and their Compositionality](https://mlsanity.com/paper/1310.4546/) (2013)
- [Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations](https://mlsanity.com/paper/1602.07332/) (2016)
- [CLEVR: A Diagnostic Dataset for Compositional Language and Elementary Visual Reasoning](https://mlsanity.com/paper/1612.06890/) (2016)
- [Show, Attend and Tell: Neural Image Caption Generation with Visual Attention](https://mlsanity.com/paper/1502.03044/) (2015)
- [Show and Tell: A Neural Image Caption Generator](https://mlsanity.com/paper/1411.4555/) (2014)
- [Compositional Visual Generation with Composable Diffusion Models](https://mlsanity.com/paper/2206.01714/) (2022)
- [When and why vision-language models behave like bags-of-words, and what to do about it?](https://mlsanity.com/paper/2210.01936/) (2022)
- [Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality](https://mlsanity.com/paper/2204.03162/) (2022)
- [Supervised Learning of Universal Sentence Representations from Natural Language Inference Data](https://mlsanity.com/paper/1705.02364/) (2017)
- [Neural Module Networks](https://mlsanity.com/paper/1511.02799/) (2015)
- [Skip-Thought Vectors](https://mlsanity.com/paper/1506.06726/) (2015)
- [CREPE: Can Vision-Language Foundation Models Reason Compositionally?](https://mlsanity.com/paper/2212.07796/) (2022)
- [Good-Enough Compositional Data Augmentation](https://mlsanity.com/paper/1904.09545/) (2019)
- [An Empirical Study and Analysis of Generalized Zero-Shot Learning for Object Recognition in the Wild](https://mlsanity.com/paper/1605.04253/) (2016)
- [Task-Driven Modular Networks for Zero-Shot Compositional Learning](https://mlsanity.com/paper/1905.05908/) (2019)
- [Learning Graph Embeddings for Compositional Zero-shot Learning](https://mlsanity.com/paper/2102.01987/) (2021)
- [Open World Compositional Zero-Shot Learning](https://mlsanity.com/paper/2101.12609/) (2021)
- [Span-based Semantic Parsing for Compositional Generalization](https://mlsanity.com/paper/2009.06040/) (2020)
- [Meaning without reference in large language models](https://mlsanity.com/paper/2208.02957/) (2022)
- [Learning to Compose Soft Prompts for Compositional Zero-Shot Learning](https://mlsanity.com/paper/2204.03574/) (2022)
- [Contrastive Learning for Image Captioning](https://mlsanity.com/paper/1710.02534/) (2017)
- [Distributional Semantics and Linguistic Theory](https://mlsanity.com/paper/1905.01896/) (2019)
- [Neurocompositional computing: From the Central Paradox of Cognition to a new generation of AI systems](https://mlsanity.com/paper/2205.01128/) (2022)
- [Independent Prototype Propagation for Zero-Shot Compositionality](https://mlsanity.com/paper/2106.00305/) (2021)
- [Compositional Video Synthesis with Action Graphs](https://mlsanity.com/paper/2006.15327/) (2020)

## Built on by (most cited first, 30 of 102)

- [ConceptFusion: Open-set Multimodal 3D Mapping](https://mlsanity.com/paper/2302.07241/) (2023)
- [Probing the 3D Awareness of Visual Foundation Models](https://mlsanity.com/paper/2404.08636/) (2024)
- [Text-to-Image Diffusion Models are Zero-Shot Classifiers](https://mlsanity.com/paper/2303.15233/) (2023)
- [Compositional Abilities Emerge Multiplicatively: Exploring Diffusion Models on a Synthetic Task](https://mlsanity.com/paper/2310.09336/) (2023)
- [Understanding the Limits of Vision Language Models Through the Lens of the Binding Problem](https://mlsanity.com/paper/2411.00238/) (2024)
- [Goal Representations for Instruction Following: A Semi-Supervised Language Interface to Control](https://mlsanity.com/paper/2307.00117/) (2023)
- [CLIP Behaves like a Bag-of-Words Model Cross-modally but not Uni-modally](https://mlsanity.com/paper/2502.03566/) (2025)
- [ImageNet3D: Towards General-Purpose Object-Level 3D Understanding](https://mlsanity.com/paper/2406.09613/) (2024)
- [Prompting Language-Informed Distribution for Compositional Zero-Shot Learning](https://mlsanity.com/paper/2305.14428/) (2023)
- [When are Lemons Purple? The Concept Association Bias of Vision-Language Models](https://mlsanity.com/paper/2212.12043/) (2022)
- [Relational Composition in Neural Networks: A Survey and Call to Action](https://mlsanity.com/paper/2407.14662/) (2024)
- [Compositional Capabilities of Autoregressive Transformers: A Study on Synthetic, Interpretable Tasks](https://mlsanity.com/paper/2311.12997/) (2023)
- [Is CLIP ideal? No. Can we fix it? Yes!](https://mlsanity.com/paper/2503.08723/) (2025)
- [When does compositional structure yield compositional generalization? A kernel theory](https://mlsanity.com/paper/2405.16391/) (2024)
- [A Philosophical Introduction to Language Models - Part II: The Way Forward](https://mlsanity.com/paper/2405.03207/) (2024)
- [Pre-trained Vision-Language Models Learn Discoverable Visual Concepts](https://mlsanity.com/paper/2404.12652/) (2024)
- [VladVA: Discriminative Fine-tuning of LVLMs](https://mlsanity.com/paper/2412.04378/) (2024)
- [Are Diffusion Models Vision-And-Language Reasoners?](https://mlsanity.com/paper/2305.16397/) (2023)
- [If CLIP Could Talk: Understanding Vision-Language Model Representations Through Their Preferred Concept Descriptions](https://mlsanity.com/paper/2403.16442/) (2024)
- [Temporal Representation Alignment: Successor Features Enable Emergent Compositionality in Robot Instruction Following](https://mlsanity.com/paper/2502.05454/) (2025)
- [A Complexity-Based Theory of Compositionality](https://mlsanity.com/paper/2410.14817/) (2024)
- [On the Limitations of Vision-Language Models in Understanding Image Transforms](https://mlsanity.com/paper/2503.09837/) (2025)
- [Visual symbolic mechanisms: Emergent symbol processing in vision language models](https://mlsanity.com/paper/2506.15871/) (2025)
- [VELOCITI: Benchmarking Video-Language Compositional Reasoning with Strict Entailment](https://mlsanity.com/paper/2406.10889/) (2024)
- [Visual Data-Type Understanding does not emerge from Scaling Vision-Language Models](https://mlsanity.com/paper/2310.08577/) (2023)
- [From Frege to chatGPT: Compositionality in language, cognition, and deep neural networks](https://mlsanity.com/paper/2405.15164/) (2024)
- [Towards Compositionality in Concept Learning](https://mlsanity.com/paper/2406.18534/) (2024)
- [Do Pre-trained Vision-Language Models Encode Object States?](https://mlsanity.com/paper/2409.10488/) (2024)
- [VSC: Visual Search Compositional Text-to-Image Diffusion Model](https://mlsanity.com/paper/2505.01104/) (2025)
- [Scaling can lead to compositional generalization](https://mlsanity.com/paper/2507.07207/) (2025)

## Abstract (arXiv)

Large-scale neural network models combining text and images have made incredible progress in recent years. However, it remains an open question to what extent such models encode compositional representations of the concepts over which they operate, such as correctly identifying "red cube" by reasoning over the constituents "red" and "cube". In this work, we focus on the ability of a large pretrained vision and language model (CLIP) to encode compositional concepts and to bind variables in a structure-sensitive way (e.g., differentiating "cube behind sphere" from "sphere behind cube"). To inspect the performance of CLIP, we compare several architectures from research on compositional distributional semantics models (CDSMs), a line of research that attempts to implement traditional compositional linguistic structures within embedding spaces. We benchmark them on three synthetic datasets - single-object, two-object, and relational - designed to test concept binding. We find that CLIP can compose concepts in a single-object setting, but in situations where concept binding is needed, performance drops dramatically. At the same time, CDSMs also perform poorly, with best performance at chance level.