The DOI matches its Crossref record; 26 of the 27 references it cites were found in a registry or an index.
How Contextual are Contextualized Word Representations? Comparing the Geometry of BERT, ELMo, and GPT-2 Embeddings
Open accessAbstract
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.
Overview
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Precedent exists for using PCA to extract network-level patterns from high-dimensional neural data, including fMRI (Smith et al., 2015), EEG (Blankertz et al., 2011), and even LLM embeddings (Ethayarajh, 2019).”
From Can "consciousness" be observed from large language model (LLM) internal states? Dissecting LLM representations obtained from Theory of Mind test with Integrated Information Theory and Span Representation analysis (2025) · cites this paper as Ethayarajh, 2019
Experimental — accuracy not yet measuredThe check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“Ethayarajh (2019) notes that unidirectional models produce embeddings that vary smoothly across tokens, implying that the difference operation can still capture meaningful transitions within a span.”
From Can "consciousness" be observed from large language model (LLM) internal states? Dissecting LLM representations obtained from Theory of Mind test with Integrated Information Theory and Span Representation analysis (2025) · cites this paper as Ethayarajh, 2019
Experimental — accuracy not yet measuredIn the citing paper
“Previous literature showed the BERT embeddings54 capture rich semantic information and are widely recognized as effective for sentence similarity tasks55–57.”
From Towards a holistic framework for multimodal LLM in 3D brain CT radiology report generation (2025) · cites this paper as [57]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173
BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding
Devlin, J., Chang, M.-W., Lee, K., & Toutanova, K. (2019) · Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers) · doi:10.18653/v1/n19-1423 · Cited by 54,325
Glove: Global Vectors for Word Representation
Pennington, J., Socher, R., & Manning, C. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1162 · Cited by 34,071
Distributed Representations of Words and Phrases and their Compositionality
Mikolov, T., Sutskever, I., Chen, K., Corrado, G., & Dean, J. (2013) · arXiv · doi:10.48550/arxiv.1310.4546 · Cited by 24,533
RoBERTa: A Robustly Optimized BERT Pretraining Approach
Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., & Stoyanov, V. (2019) · arXiv · doi:10.48550/arxiv.1907.11692 · Cited by 16,515
Deep Contextualized Word Representations
Peters, M., Neumann, M., Iyyer, M., Gardner, M., Clark, C., Lee, K., & Zettlemoyer, L. (2018) · Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers) · doi:10.18653/v1/n18-1202 · Cited by 11,983
XLNet: Generalized Autoregressive Pretraining for Language Understanding
Yang, Z., Dai, Z., Yang, Y., Carbonell, J., Salakhutdinov, R., & Le, Q. V. (2019) · arXiv · doi:10.48550/arxiv.1906.08237 · Cited by 6,687
How transferable are features in deep neural networks?
Yosinski, J., Clune, J., Bengio, Y., & Lipson, H. (2014) · arXiv · doi:10.48550/arxiv.1411.1792 · Cited by 6,061
Linguistic Regularities in Continuous Space Word Representations
Mikolov, T., Yih, W.-T., & Zweig, G. (2013) · Cited by 2,913
Neural Word Embedding as Implicit Matrix Factorization
Levy, O., & Goldberg, Y. (2014) · neural information processing systems · Cited by 1,581
Research works printed but not linked: 5
- Eneko Agirre, Daniel Cer, Mona Diab, Aitor Gonzalez- Agirre, and Weiwei Guo. 2013. Sem 2013 shared task: Semantic textual similarity, including a pilot on typed-similarity. In SEM 2013: The Second Joint Conference on Lexical and Computational Seman- tics. Association for Computational Linguistics.
- John Hewitt and Christopher D. Manning. 2019. A structural probe for finding syntax in word represen- tations. In North American Chapter of the Associ- ation for Computational Linguistics: Human Lan- guage Technologies. Association for Computational Linguistics.
- Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S Cor- rado, and Jeff Dean. 2013a. Distributed representa- tions of words and phrases and their compositional- ity. In Advances in Neural Information Processing Systems. pages 3111-3119.
- Matthew Peters, Mark Neumann, Mohit Iyyer, Matt Gardner, Christopher Clark, Kenton Lee, and Luke Zettlemoyer. 2018. Deep contextualized word rep- resentations. In Proceedings of the 2018 Confer- ence of the North American Chapter of the Associ- ation for Computational Linguistics: Human Lan- guage Technologies, Volume 1 (Long Papers). pages 2227-2237.
- Ian Tenney, Patrick Xia, Berlin Chen, Alex Wang, Adam Poliak, R. Thomas McCoy, Najoung Kim, Benjamin Van Durme, Samuel R. Bowman, Dipan- jan Das, and Ellie Pavlick. 2019. What do you learn from context? probing for sentence structure in contextualized word representations. In Inter- national Conference on Learning Representations. https://openreview.net/forum?id=SJzSgnRcKX.
Cited by
Sorted by citations
SimCSE: Simple Contrastive Learning of Sentence Embeddings
Gao, T., Yao, X., & Chen, D. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.552 · Cited by 2,851
A Metaverse: Taxonomy, Components, Applications, and Open Challenges
Park, S., & Kim, Y. (2022) · IEEE Access · doi:10.1109/access.2021.3140175 · Cited by 2,046
A survey on sentiment analysis methods, applications, and challenges
Wankhade, M., Rao, A. C. S., & Kulkarni, C. (2022) · Artificial Intelligence Review · doi:10.1007/s10462-022-10144-1 · Cited by 1,892
A primer in BERTology: What we know about how BERT works
Rogers, A., Kovaleva, O., & Rumshisky, A. (2020) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00349 · Cited by 1,552
Pre-trained models: Past, present and future
Han, X., Zhang, Z., Ding, N., Gu, Y., Liu, X., Huo, Y., Qiu, J., Yao, Y., Zhang, A., Zhang, L., et al. (2021) · AI Open · doi:10.1016/j.aiopen.2021.08.002 · Cited by 994
Metadata
- Authors
- Ethayarajh, K.
- Affiliations
- Stanford University (US)
- Corresponding author
- Ethayarajh, Kawin
- Venue
- Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)
- Published
- 2019
- Pages
- 55-65
- OpenAlex
- W2971569798
- Cited by
- 841
- References
- 27
- Citations by year
- 2026: 93 · 2025: 92 · 2024: 115 · 2023: 160 · 2022: 158 · 2021: 157 · 2020: 64
- FWCI
- 36.26
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Conference on Empirical Methods in Natural Language Processing (EMNLP) (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- Natural Sciences and Engineering Research Council of Canada
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Natural Language Processing Techniques
- Physical Sciences › Computer Science › Artificial Intelligence › Explainable Artificial Intelligence (XAI)
- Keywords
- contextualized word representations, word embeddings, BERT, ELMo, GPT-2, embedding geometry, cosine similarity, static embeddings, layer-wise analysis
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
