Riemannian Geometry for Pre-trained Language Model Embeddings
cs.CL, cs.AI
Submitted: 2026-07-08
Updated: 2026-09-21
Code: https://github.com/hotherio/intrinsic-green-learning
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- The largest EEG-based BCI reproducibility study for open science: the MOABB benchmark
- Truth as a Trajectory: What Internal Representations Reveal About Large Language Model Reasoning
- The Curved Spacetime of Transformer Architectures
- Score-based Pullback Riemannian Geometry: Extracting the Data Manifold Geometry using Anisotropic Flows
- Emergent Riemannian geometry over learning discrete computations on continuous manifolds
- RiemannFormer: A Framework for Attention in Curved Spaces
- SPD Matrix Learning for Neuroimaging Analysis: Perspectives, Methods, and Challenges
- Measuring Intrinsic Dimension of Token Embeddings
- Curvature-aware Manifold Learning
- Curved Inference: Concern-Sensitive Geometry in Large Language Model Residual Streams
- UMAP: Uniform Manifold Approximation and Projection for Dimension Reduction
- The Geometry of Truth: Layer-wise Semantic Dynamics for Hallucination Detection in Large Language Models
- Local Intrinsic Dimension Unveils Hallucinations in Diffusion Models
- Mechanistic Decomposition of Sentence Representations
- Neural Network Acceptability Judgments
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering