mlsanityData to 7 Oct 2026

The Platonic Representation Hypothesis

arXiv 2405.07987 ↗First posted 13 May 2024cs.LG cs.AI cs.CVcited by 362

Not on the map. Why?

Cited by 38 in Multimodal LLMs / vision-language models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

We argue that representations in AI models, particularly deep networks, are converging. First, we survey many examples of convergence in the literature: over time and across multiple domains, the ways by which different neural networks represent data are becoming more aligned. Next, we demonstrate convergence across data modalities: as vision models and language models get larger, they measure distance between datapoints in a more and more alike way. We hypothesize that this convergence is driving toward a shared statistical model of reality, akin to Plato's concept of an ideal reality. We term such a representation the platonic representation and discuss several possible selective pressures toward it. Finally, we discuss the implications of these trends, their limitations, and counterexamples to our analysis.

Key facts. Cited by 362 papers in the map (data through 7 Oct 2026). Builds on 85 of its 144 references that are in the map. First posted on arXiv 13 May 2024.

Builds on

85 of 144 references.

References · 85

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  3. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  4. 2023GPT-4 Technical Reportcited by 18,132
  5. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  6. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 12,865
  7. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  8. 2018The Unreasonable Effectiveness of Deep Features as a Perceptual Metriccited by 11,029
  9. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  10. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,441
  11. 2023Visual Instruction Tuningcited by 7,792
  12. 2018Representation Learning with Contrastive Predictive Codingcited by 7,662
38 more
  1. 2021Masked Autoencoders Are Scalable Vision Learnerscited by 7,339
  2. 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,893
  3. 2020Scaling Laws for Neural Language Modelscited by 6,411
  4. 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,018
  5. 2021On the Opportunities and Risks of Foundation Modelscited by 3,961
  6. 2023Mistral 7Bcited by 3,065
  7. 2023RT-2: Vision-Language-Action Models Transfer Web Knowledge to Robotic Controlcited by 2,993
  8. 2018UMAP: Uniform Manifold Approximation and Projection for Dimension Reductioncited by 2,841
  9. 2019HuggingFace's Transformers: State-of-the-art Natural Language Processingcited by 2,201
  10. 2021SimCSE: Simple Contrastive Learning of Sentence Embeddingscited by 2,168
  11. 2014Conditional Generative Adversarial Netscited by 2,152
  12. 2023PaLM-E: An Embodied Multimodal Language Modelcited by 2,037
  13. 2022Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language modelscited by 1,970
  14. 2022BLOOM: A 176B-Parameter Open-Access Multilingual Language Modelcited by 1,968
  15. 2024Mixtral of Expertscited by 1,785
  16. 2019Similarity of Neural Network Representations Revisitedcited by 1,695
  17. 2018World Modelscited by 1,444
  18. 2020Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hyperspherecited by 1,393
  19. 2022Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference timecited by 1,366
  20. 2019Contrastive Multiview Codingcited by 1,297
  21. 2020Prevalence of Neural Collapse during the terminal phase of deep learning trainingcited by 713
  22. 2024OLMo: Accelerating the Science of Language Modelscited by 599
  23. 2018Loss Surfaces, Mode Connectivity, and Fast Ensembling of DNNscited by 585
  24. 2017Deep Learning Scaling is Predictable, Empiricallycited by 556
  25. 2019A Theoretical Analysis of Contrastive Unsupervised Representation Learningcited by 530
  26. 2022StyleGAN-XL: Scaling StyleGAN to Large Diverse Datasetscited by 489
  27. 2022Git Re-Basin: Merging Models modulo Permutation Symmetriescited by 460
  28. 2020Rethinking Few-Shot Image Classification: a Good Embedding Is All You Need?cited by 435
  29. 2019Implicit Regularization in Deep Matrix Factorizationcited by 402
  30. 2021WIT: Wikipedia-based Image Text Dataset for Multimodal Multilingual Machine Learningcited by 285
  31. 2018Implicit Bias of Gradient Descent on Linear Convolutional Networkscited by 251
  32. 2023LLM-grounded Diffusion: Enhancing Prompt Understanding of Text-to-Image Diffusion Models with Large Language Modelscited by 242
  33. 2014Understanding image representations by measuring their equivariance and equivalencecited by 240
  34. 2019Uniform convergence may be unable to explain generalization in deep learningcited by 223
  35. 2021Contrastive Learning Inverts the Data Generating Processcited by 220
  36. 2021Revisiting Model Stitching to Compare Neural Representationscited by 203
  37. 2023Large Language Models as General Pattern Machinescited by 196
  38. 2018Deep learning generalizes because the parameter-function map is biased towards simple functionscited by 180

50 most cited of 85 All 85 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 62; 2025: 147; 2026: 153.

362 papers.

Sort:

By area:

  1. 2024Representation Alignment for Generation: Training Diffusion Transformers Is Easier Than You Thinkcited by 666
  2. 2025REPA-E: Unlocking VAE for End-to-End Tuning with Latent Diffusion Transformerscited by 212
  3. 2024MetaMorph: Multimodal Understanding and Generation via Instruction Tuningcited by 205
  4. 2025RL's Razor: Why Online Reinforcement Learning Forgets Lesscited by 162
  5. 2024Lumina-mGPT: Illuminate Flexible Photorealistic Text-to-Image Generation with Multimodal Generative Pretrainingcited by 148
  6. 2024AuroraCap: Efficient, Performant Video Detailed Captioning and a New Benchmarkcited by 137
  7. 2024Analyzing the Generalization and Reliability of Steering Vectorscited by 130
  8. 2025Representation Entanglement for Generation: Training Diffusion Transformers Is Much Easier Than You Thinkcited by 95
  9. 2024Truth is Universal: Robust Detection of Lies in LLMscited by 78
  10. 2024Dual Diffusion for Unified Image Generation and Understandingcited by 76
  11. 2024The Semantic Hub Hypothesis: Language Models Share Semantic Representations Across Languages and Modalitiescited by 69
  12. 2025Universal Sparse Autoencoders: Interpretable Cross-Model Concept Alignmentcited by 62
38 more
  1. 2025Scaling Language-Free Visual Representation Learningcited by 62
  2. 2025Interpreting CLIP with Hierarchical Sparse Autoencoderscited by 54
  3. 2025Harnessing the Universal Geometry of Embeddingscited by 54
  4. 2024Knowledge Mechanisms in Large Language Models: A Survey and Perspectivecited by 50
  5. 2025How Linguistics Learned to Stop Worrying and Love the Language Modelscited by 50
  6. 2025Visual Representation Alignment for Multimodal Large Language Modelscited by 40
  7. 2024Align Anything: Training All-Modality Models to Follow Instructions with Language Feedbackcited by 38
  8. 2025AlignTok: Aligning Visual Foundation Encoders to Tokenizers for Diffusion Modelscited by 35
  9. 2025Shared Global and Local Geometry of Language Model Embeddingscited by 34
  10. 2025TUNA: Taming Unified Visual Representations for Native Unified Multimodal Modelscited by 34
  11. 2024Towards Universality: Studying Mechanistic Similarity Across Language Model Architecturescited by 31
  12. 2024Assessing and Learning Alignment of Unimodal Vision and Language Modelscited by 31
  13. 2024Unifying Causal Representation Learning with the Invariance Principlecited by 30
  14. 2024Cross-Entropy Is All You Need To Invert the Data Generating Processcited by 30
  15. 2026LongCat-Next: Lexicalizing Modalities as Discrete Tokenscited by 30
  16. 2025Understanding the Emergence of Multimodal Representation Alignmentcited by 28
  17. 2025Learning to See Before Seeing: Demystifying LLM Visual Priors from Language Pre-trainingcited by 23
  18. 2025Cycle Consistency as Reward: Learning Image-Text Alignment without Human Preferencescited by 22
  19. 2024Automating the Search for Artificial Life with Foundation Modelscited by 20
  20. 2025Questioning Representational Optimism in Deep Learning: The Fractured Entangled Representation Hypothesiscited by 19
  21. 2026Visual Generation Unlocks Human-Like Reasoning through Multimodal World Modelscited by 18
  22. 2025Transferring Linear Features Across Language Models With Model Stitchingcited by 17
  23. 2024Revisiting In-context Learning Inference Circuit in Large Language Modelscited by 16
  24. 2025From superposition to sparse codes: interpretable representations in neural networkscited by 16
  25. 2025Skywork-R1V3 Technical Reportcited by 16
  26. 2025Beyond Token Probes: Hallucination Detection via Activation Tensors with ACT-ViTcited by 16
  27. 2025Contrastive Representation Regularization for Vision-Language-Action Modelscited by 16
  28. 2025VFM-VAE: Vision Foundation Models Can Be Good Tokenizers for Latent Diffusion Modelscited by 16
  29. 2024See What LLMs Cannot Answer: A Self-Challenge Framework for Uncovering LLM Weaknessescited by 15
  30. 2024Vision-Language Models Create Cross-Modal Task Representationscited by 14
  31. 2024An Information Criterion for Controlled Disentanglement of Multimodal Datacited by 14
  32. 2024Neural Scaling Laws Rooted in the Data Distributioncited by 14
  33. 2024Universal dimensions of visual representationcited by 12
  34. 2025Aligning Multimodal Representations through an Information Bottleneckcited by 12
  35. 2025Learning Diffusion Models with Flexible Representation Guidancecited by 12
  36. 2025CauKer: Classification Time Series Foundation Models Can Be Pretrained on Synthetic Datacited by 12
  37. 2025Disentangling the Factors of Convergence between Brains and Computer Vision Modelscited by 12
  38. 2025To Align or Not to Align: Strategic Multimodal Representation Alignment for Optimal Performancecited by 12

50 most cited of 362 All 362 (JSON)