A Hub of Short Rows Inflates Intrinsic Dimension Estimation of Token Embeddings

arXiv:2608.29702 · cs.CL · Submitted 2026-08-30 · Read on arXiv

cs.CL

Submitted: 2026-08-30

Updated: 2026-08-30

Comments: Submitted to NeurReps Workshop @ NeurIPS 2026

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Intrinsic Dimension — 24× in this paper · explained in 2 episodes
• Embeddings — 13× in this paper · explained in 2 episodes
• Empirical Methods in Natural Language Processing — 3× in this paper
• Transformer — 3× in this paper · explained in 2 episodes
• Association for Computational Linguistics — 2× in this paper
• Calibration — 2× in this paper · explained in 7 episodes
• Conference on Knowledge Discovery and Data Mining — 2× in this paper
• In International Conference on Learning Representations — 2× in this paper
• Redundancy — 2× in this paper · explained in 2 episodes
• AI Alignment — 1× in this paper · explained in 2 episodes
• Alignment — 1× in this paper · explained in 1 episode
• Ambient Dimension — 1× in this paper · explained in 1 episode
• Anisotropy — 1× in this paper · explained in 3 episodes
• Fine-Tuning — 1× in this paper · explained in 5 episodes
• Robustness — 1× in this paper · explained in 5 episodes
• Trace — 1× in this paper · explained in 2 episodes

Sources

• Intrinsic Dimensionality Explains the Effectiveness of Language Model Fine-Tuning
• Improving zero-shot learning by mitigating the hubness problem
• Representation Degeneration Problem in Training Natural Language Generation Models
• Measuring Intrinsic Dimension of Token Embeddings
• All-but-the-Top: Simple and Effective Postprocessing for Word Representations
• The Shape of Learning: Anisotropy and Intrinsic Dimensions in Transformer-Based Models
• Token embeddings violate the manifold hypothesis
• Your UnEmbedding Matrix is Secretly a Feature Lens for Text Embeddings

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