# Back into Plato's Cave: Examining Cross-modal Representational Convergence at Scale

arXiv: https://arxiv.org/abs/2604.18572 · HTML: https://mlsanity.com/paper/2604.18572/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- Builds on 73 of its 112 references that are in the map.
- First posted on arXiv 20 Apr 2026.

## Builds on (most cited first, 30 of 73)

- [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)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Training Verifiers to Solve Math Word Problems](https://mlsanity.com/paper/2110.14168/) (2021)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Measuring Massive Multitask Language Understanding](https://mlsanity.com/paper/2009.03300/) (2020)
- [Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challenge](https://mlsanity.com/paper/1803.05457/) (2018)
- [HellaSwag: Can a Machine Really Finish Your Sentence?](https://mlsanity.com/paper/1905.07830/) (2019)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [Qwen3-VL Technical Report](https://mlsanity.com/paper/2511.21631/) (2025)
- [Visual Genome: Connecting Language and Vision Using Crowdsourced Dense Image Annotations](https://mlsanity.com/paper/1602.07332/) (2016)
- [Do As I Can, Not As I Say: Grounding Language in Robotic Affordances](https://mlsanity.com/paper/2204.01691/) (2022)
- [Gemma 2: Improving Open Language Models at a Practical Size](https://mlsanity.com/paper/2408.00118/) (2024)
- [Gemma 3 Technical Report](https://mlsanity.com/paper/2503.19786/) (2025)
- [Mixtral of Experts](https://mlsanity.com/paper/2401.04088/) (2024)
- [Similarity of Neural Network Representations Revisited](https://mlsanity.com/paper/1905.00414/) (2019)
- [LAION-400M: Open Dataset of CLIP-Filtered 400 Million Image-Text Pairs](https://mlsanity.com/paper/2111.02114/) (2021)
- [DINOv3](https://mlsanity.com/paper/2508.10104/) (2025)
- [Code as Policies: Language Model Programs for Embodied Control](https://mlsanity.com/paper/2209.07753/) (2022)
- [Gemma: Open Models Based on Gemini Research and Technology](https://mlsanity.com/paper/2403.08295/) (2024)
- [Large-scale Contrastive Language-Audio Pretraining with Feature Fusion and Keyword-to-Caption Augmentation](https://mlsanity.com/paper/2211.06687/) (2022)
- [Yi: Open Foundation Models by 01.AI](https://mlsanity.com/paper/2403.04652/) (2024)
- [Phi-4 Technical Report](https://mlsanity.com/paper/2412.08905/) (2024)
- [FLAVA: A Foundational Language And Vision Alignment Model](https://mlsanity.com/paper/2112.04482/) (2021)
- [OLMo: Accelerating the Science of Language Models](https://mlsanity.com/paper/2402.00838/) (2024)
- [Language Is Not All You Need: Aligning Perception with Language Models](https://mlsanity.com/paper/2302.14045/) (2023)

## Built on by (most cited first, 16 of 16)

- [The Capability Convergence Hypothesis: Capability from Access Structure, Not Scale](https://mlsanity.com/paper/2607.14144/) (2026)
- [Better Together: Evaluating the Complementarity of Earth Embedding Models](https://mlsanity.com/paper/2605.18667/) (2026)
- [A Unifying Framework for Concept-Based Representational Similarity](https://mlsanity.com/paper/2606.09653/) (2026)
- [PlatonicNav: Unveiling Semantic Correspondence in Navigation with Platonic Topological Maps](https://mlsanity.com/paper/2606.01788/) (2026)
- [The Cross-Architecture Substrate: A Domain-Transcendent, Calibration-Surviving Geometric Invariant of Modern Vision Encoders](https://mlsanity.com/paper/2606.07882/) (2026)
- [The Platonic Defense: Backdoor Defense for Self-Supervised Encoders in the Era of Large Scale Pre-training](https://mlsanity.com/paper/2606.29451/) (2026)
- [Medical foundation models converge less under label supervision](https://mlsanity.com/paper/2607.20274/) (2026)
- [Cross-Lingual Alignment Without Joint Training: Do Monolingual Language Models Converge on Universal Representations?](https://mlsanity.com/paper/2608.27115/) (2026)
- [SharedSAE: One Feature Dictionary Across Language Models](https://mlsanity.com/paper/2609.04344/) (2026)
- [The Platonic brain bridge hypothesis: human brain networks as an architectural prior for multimodal large language models](https://mlsanity.com/paper/2609.10947/) (2026)
- [Displacement Geometry Captures Platonic Shared Reality Across Models and Modalities](https://mlsanity.com/paper/2609.24209/) (2026)
- [Spherical Interpolation for Backward-Compatible Multimodal Representations](https://mlsanity.com/paper/2609.39836/) (2026)
- [Shared Geometry Is Not Shared Physics: A Layerwise Test of the Platonic Representation Hypothesis in Astronomy](https://mlsanity.com/paper/2610.04130/) (2026)
- [A theory of platonic representations in language models](https://mlsanity.com/paper/2610.07168/) (2026)
- [Shared Geometry As A Rosetta Stone: Cross-Modal Alignment Without Paired Data](https://mlsanity.com/paper/2610.09411/) (2026)
- [Vision Transformers Learn Word Meaning from Images Alone](https://mlsanity.com/paper/2610.13252/) (2026)

## Abstract (arXiv)

The Platonic Representation Hypothesis posits that neural networks trained on different modalities (e.g., text and images) converge toward a shared representation of reality. If true, this has significant implications for whether modality choice matters at all. In this paper, we show that the evidence for this claim is substantially weaker than subsequent work suggests. The mutual k-nearest-neighbor metric used on 1024 text-image pairs in the original study captures only coarse structure. To keep the alignment from collapsing as one scales up the data, k has to grow proportionally, undercutting the argument for fine-grained representational convergence. The reported increase in alignment with language model strength saturates for recent models. Moreover, the one-to-one text-image pairing favors alignment, while alignment decreases with non-bijective data. We further find that image and text representations indeed share coarse semantic structure, but neither stronger language models nor richer captions yield fine-grained alignment. Thus, multimodal representations share coarse structure without evidence of convergence to a shared representation -- arguably, full representational convergence would require fine-grained alignment.