Displacement Geometry Captures Platonic Shared Reality Across Models and Modalities
cs.LG
Submitted: 2026-09-21
Updated: 2026-09-21
Code: https://github.com/HelloSCM/Representation_Displacements
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- LLM2Vec: Large Language Models Are Secretly Powerful Text Encoders
- LLM2Vec-Gen: Generative Embeddings from Large Language Models
- Toy Models of Superposition
- Revisiting the Platonic Representation Hypothesis: An Aristotelian View
- Better Together: Leveraging Unpaired Multimodal Data for Stronger Unimodal Models
- Distilling the Knowledge in a Neural Network
- Modulating Cross-Modal Convergence with Single-Stimulus, Intra-Modal Dispersion
- LLM2CLIP: Powerful Language Model Unlocks Richer Cross-Modality Representation
- The Universal Weight Subspace Hypothesis
- Back into Plato's Cave: Examining Cross-modal Representational Convergence at Scale
- Relative representations enable zero-shot latent space communication
- The Linear Representation Hypothesis and the Geometry of Large Language Models
- The Geometry of Categorical and Hierarchical Concepts in Large Language Models
- Fantastic Gains and Where to Find Them: On the Existence and Prospect of General Knowledge Transfer between Any Pretrained Model
- Better Language Models Exhibit Higher Visual Alignment
- GLU Variants Improve Transformer
- Getting aligned on representational alignment
- Understanding the Emergence of Multimodal Representation Alignment
- Compositional Generalization Requires Linear, Orthogonal Representations in Vision Embedding Models
- Words That Make Language Models Perceive
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks