Pith. sign in

REVIEW 20 cited by

How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 1909.00512 v1 pith:5BCQ7KQR submitted 2019-09-02 cs.CL

How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings

classification cs.CL
keywords representationswordcontextualizedlayersbertelmoupperwords
verification ladder T0 review T1 audit T2 compute T3 formal T4 reserved
0 comments
read the original abstract

Replacing static word embeddings with contextualized word representations has yielded significant improvements on many NLP tasks. However, just how contextual are the contextualized representations produced by models such as ELMo and BERT? Are there infinitely many context-specific representations for each word, or are words essentially assigned one of a finite number of word-sense representations? For one, we find that the contextualized representations of all words are not isotropic in any layer of the contextualizing model. While representations of the same word in different contexts still have a greater cosine similarity than those of two different words, this self-similarity is much lower in upper layers. This suggests that upper layers of contextualizing models produce more context-specific representations, much like how upper layers of LSTMs produce more task-specific representations. In all layers of ELMo, BERT, and GPT-2, on average, less than 5% of the variance in a word's contextualized representations can be explained by a static embedding for that word, providing some justification for the success of contextualized representations.

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Forward citations

Cited by 20 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

  1. REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations

    cs.CL 2026-05 unverdicted novelty 8.0

    REALISTA optimizes continuous combinations of valid editing directions in latent space to produce realistic adversarial prompts that elicit hallucinations more effectively than prior methods, including on large reason...

  2. Channel Location Constrains the Auditability of Subliminal Learning

    cs.LG 2026-06 unverdicted novelty 7.0

    Auditability of subliminal learning is constrained by channel location, with initialization-dependent body channels allowing pre-training screens while vocabulary geometry and conditional body channels evade them.

  3. ResRL: Boosting LLM Reasoning via Negative Sample Projection Residual Reinforcement Learning

    cs.LG 2026-05 unverdicted novelty 7.0

    ResRL decouples shared semantics between positive and negative responses in LLM reinforcement learning via SVD-based projection residuals, outperforming baselines including NSR by up to 9.4% on math reasoning benchmarks.

  4. mEOL: Training-Free Instruction-Guided Multimodal Embedder for Vector Graphics and Image Retrieval

    cs.CV 2026-04 unverdicted novelty 7.0

    mEOL creates aligned embeddings for text, images, and SVGs using instruction-guided MLLM one-word summaries and semantic SVG rewriting, outperforming baselines on a new text-to-SVG retrieval benchmark.

  5. Riemannian Geometry for Pre-trained Language Model Embeddings

    cs.CL 2026-07 conditional novelty 6.0

    Aggregating per-token pullback metrics via the Fréchet mean on the SPD manifold outperforms Euclidean mean pooling for sentence classification, with most of the gain attributable to geometric aggregation rather than l...

  6. Decoupled Residual Quantization for Robust Semantic IDs in Recommendation

    cs.IR 2026-06 unverdicted novelty 6.0

    Presents a diagnostic framework for semantic ID tokenizer failures using overlap and capacity metrics and proposes DRQ to separate geometry from distribution matching.

  7. NITP: Next Implicit Token Prediction for LLM Pre-training

    cs.CL 2026-05 unverdicted novelty 6.0

    NITP augments standard next-token prediction with implicit semantic prediction in representation space using shallow-layer self-supervision, reporting consistent downstream gains on 0.5B-9B models including 5.7% on MM...

  8. HyperPersona: A Multi-Level Hypergraph Framework for Text-Based Automatic Personality Prediction

    cs.AI 2026-05 unverdicted novelty 6.0

    HyperPersona is a hypergraph framework that jointly models document, sentence, and word levels of text via hyperedges and nodes, then uses a transformer graph encoder to predict Big Five personality traits from text alone.

  9. REALISTA: Realistic Latent Adversarial Attacks that Elicit LLM Hallucinations

    cs.CL 2026-05 unverdicted novelty 6.0

    REALISTA generates semantically coherent adversarial prompts via latent-space optimization over input-dependent editing directions, achieving stronger hallucination elicitation than prior realistic attacks on open-sou...

  10. ResRL: Boosting LLM Reasoning via Negative Sample Projection Residual Reinforcement Learning

    cs.LG 2026-05 unverdicted novelty 6.0

    ResRL boosts LLM reasoning by modulating negative gradients with SVD-based projection residuals from negative samples, outperforming NSR by 9.4% Avg@16 on math benchmarks while preserving diversity across 12 tasks.

  11. Geometry-Calibrated Conformal Abstention for Language Models

    cs.CL 2026-04 unverdicted novelty 6.0

    Geometry-calibrated conformal abstention lets language models abstain from uncertain queries with finite-sample guarantees on both participation rate and conditional correctness of answers.

  12. REZE: Representation Regularization for Domain-adaptive Text Embedding Pre-finetuning

    cs.CL 2026-04 unverdicted novelty 6.0

    REZE controls representation shifts in contrastive pre-finetuning of text embeddings via eigenspace decomposition of anchor-positive pairs and adaptive soft-shrinkage on task-variant directions.

  13. GPC: Large-Scale Generative Pretraining for Transferable Motor Control

    cs.CV 2026-06 unverdicted novelty 5.0

    GPC learns a motion vocabulary via Finite Scalar Quantization and end-to-end RL, then trains an autoregressive transformer for next-token control generation, achieving 99.98% motion reproduction success with emergent ...

  14. The Crowded Embedding Space: A Mean-Field Mechanism for Emergent Marginalization in Retrieval-Augmented Agents

    cs.IR 2026-06 unverdicted novelty 5.0

    A mean-field analysis of embedding-space crowding shows a phase transition and Fokker-Planck dynamics that drive retrieval-augmented agents to self-organize toward exclusive service of majority interests.

  15. When Meaning Travels: A Granular Lens on Hybrid-MoE's Role in Idiomatic Understanding for Language Models

    cs.CL 2026-06 unverdicted novelty 5.0

    HybridMoE with controlled hybridization and idiomatic property signals yields 5-6% gains in figurative language representation for multilingual vision-language models.

  16. Improving Relative Representations with Learned Anchors and Whitened Inner Products

    cs.LG 2026-05 unverdicted novelty 5.0

    Learned anchors as semantic prototypes combined with whitened inner products improve relative representations, enabling nearly lossless zero-shot communication between heterogeneous neural models on vision and language tasks.

  17. NITP: Next Implicit Token Prediction for LLM Pre-training

    cs.CL 2026-05 unverdicted novelty 5.0

    NITP adds dense supervision from shallow model layers to predict implicit next-token semantics, yielding consistent downstream gains on 0.5B-9B models with ~2% extra training FLOPs.

  18. Representation Without Reward: A JEPA Audit for LLM Fine-Tuning

    cs.LG 2026-05 conditional novelty 5.0

    An empirical audit of 22 JEPA-style training auxiliaries on Llama-3.2-1B fine-tuning for regex generation finds no statistically significant task improvement after multiple-testing correction, even when auxiliaries vi...

  19. HyperLens: Quantifying Cognitive Effort in LLMs with Fine-grained Confidence Trajectory

    cs.AI 2026-05 unverdicted novelty 5.0

    HyperLens reveals that deeper transformer layers magnify small confidence changes into fine-grained trajectories, allowing quantification of cognitive effort where complex tasks demand more and standard SFT can reduce it.

  20. The Hitchhiker's Guide to Agentic AI: From Foundations to Systems

    cs.AI 2026-06 unverdicted novelty 2.0

    A comprehensive reference book organizing existing techniques for agentic AI systems across LLM substrate, reasoning, agent design patterns, inter-agent coordination, and production deployment.