Research map / Label-efficient and robust learning / Recognition and long-tail
Revisiting the Platonic Representation Hypothesis: An Aristotelian View
Fabian Gröger, Shuo Wen, Maria Brbic
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add null calibration to linear CKA | Linear CKA · Image-text convergence (WIT, 204 pairs) | Pearson correlation with LM ranking: 0.86 -> 0.45 | Hurt |
| add null calibration to Procrustes distance | Procrustes distance · Image-text convergence (WIT, 204 pairs) | Pearson correlation with LM ranking: 0.89 -> 0.39 | Hurt |
| add null calibration to mutual kNN and CKNNA | Local neighborhood metrics · Image-text convergence (WIT, 204 pairs) | Pearson correlation with LM ranking: mutual kNN stays near 0.85, CKNNA near 0.87 | No clear effect |
| increase layer count from 2 to 128 without aggregation-aware calibration (ablation) | Max-aggregated CKA under null · Synthetic depth experiment | max CKA score | Hurt |
| add aggregation-aware calibration instead of entry-wise calibration (ablation) | Max-aggregated CKA under null · Synthetic depth experiment | max CKA score | Improved |
| add permutation null calibration | Similarity metrics under H0 · Synthetic width experiment | Type-I error rate: rejection rates at or below alpha=0.05 | Improved |
Closest papers by meaning
- Position: The Platonic Representation Hypothesis (ICML 2024 · similarity 0.93)
- Displacement Geometry Captures Platonic Shared Reality Across Models and Modalities (NeurIPS 2026 · similarity 0.90)
- Objective drives the consistency of representational similarity across datasets (ICML 2025 · similarity 0.89)
- Platonic Representations in the Human Brain: Unsupervised Recovery of Universal Geometry (NeurIPS 2026 · similarity 0.89)
- Bridging Critical Gaps in Convergent Learning: How Representational Alignment Evolves Across Layers, Training, and Distribution Shifts (NeurIPS 2025 · similarity 0.89)
- Not all solutions are created equal: An analytical dissociation of functional and representational similarity in deep linear neural networks (ICML 2025 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
