On the Sentence Embeddings from Pre-trained Language Models
Open accessAbstract
Pre-trained contextual representations like BERT have achieved great success in natural language processing.However, the sentence embeddings from the pre-trained language models without fine-tuning have been found to poorly capture semantic meaning of sentences.In this paper, we argue that the semantic information in the BERT embeddings is not fully exploited.We first reveal the theoretical connection between the masked language model pre-training objective and the semantic similarity task theoretically, and then analyze the BERT sentence embeddings empirically.We find that BERT always induces a non-smooth anisotropic semantic space of sentences, which harms its performance of semantic similarity.To address this issue, we propose to transform the anisotropic sentence embedding distribution to a smooth and isotropic Gaussian distribution through normalizing flows that are learned with an unsupervised objective.Experimental results show that our proposed BERT-flow method obtains significant performance gains over the state-of-the-art sentence embeddings on a variety of semantic textual similarity tasks.The code is available at https://github.com/ bohanli/BERT-flow.
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
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?
BERT embeddings capture rich semantic information and are widely recognized as effective for sentence similarity tasks.
“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 [56]
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Cites
Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017) · arXiv · doi:10.48550/arxiv.1706.03762 · Cited by 120,404 · see citing works
Jakubův, J., Chvalovský, K., Goertzel, Z., et al. (2023) · DROPS (Schloss Dagstuhl – Leibniz Center for Informatics) · doi:10.4230/lipics.itp.2023.19 · Cited by 76,648 · see citing works
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 48,293 · see citing works
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 · see citing works
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 · see citing works
Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks
Reimers, N., & Gurevych, I. (2019) · Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) · doi:10.18653/v1/d19-1410 · Cited by 12,778 · see citing works
SQuAD: 100,000+ Questions for Machine Comprehension of Text
Rajpurkar, P., Zhang, J., Lopyrev, K., & Liang, P. (2016) · Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d16-1264 · Cited by 6,208 · see citing works
GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding
Wang, A., Singh, A., Michael, J., Hill, F., Levy, O., & Bowman, S. (2018) · Proceedings of the 2018 EMNLP Workshop BlackboxNLP: Analyzing and Interpreting Neural Networks for NLP · doi:10.18653/v1/w18-5446 · Cited by 4,222 · see citing works
A large annotated corpus for learning natural language inference
Bowman, S. R., Angeli, G., Potts, C., & Manning, C. D. (2015) · Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d15-1075 · Cited by 3,377 · see citing works
A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference
Williams, A., Nangia, N., & Bowman, S. (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/n18-1101 · Cited by 3,342 · see citing works
(2018) · International Conference on Learning Representations · Cited by 1,621 · see citing works
Neural Word Embedding as Implicit Matrix Factorization
Levy, O., & Goldberg, Y. (2014) · neural information processing systems · Cited by 1,581 · see citing works
NICE: Non-linear Independent Components Estimation
Dinh, L., Krueger, D., & Bengio, Y. (2014) · arXiv · doi:10.48550/arxiv.1410.8516 · Cited by 1,493 · see citing works
Cer, D., Yang, Y., Kong, S.-Y., Hua, N., Limtiaco, N., John, R. S., Constant, N., Guajardo-Cespedes, M., Yuan, S., Tar, C., et al. (2018) · arXiv · doi:10.48550/arxiv.1803.11175 · Cited by 1,267 · see citing works
Self-citation(2017) · International Conference on Learning Representations · Cited by 1,075 · see citing works
Ethayarajh, K. (2019) · Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP) · doi:10.18653/v1/d19-1006 · Cited by 841 · see citing works
A SICK cure for the evaluation of compositional distributional semantic models
Marelli, M., Menini, S., Baroni, M., Bentivogli, L., Raffaella Bernardi, B., & Zamparelli, R. (2014) · Language Resources and Evaluation Conference (LREC) · doi:10.63317/39qdhuevzbqa · Cited by 680 · see citing works
SemEval-2012 Task 6: A Pilot on Semantic Textual Similarity
Agirre, E., Cer, D. M., Diab, M., & González-Agirre, A. (2012) · Cited by 653 · see citing works
SemEval-2016 Task 1: Semantic Textual Similarity, Monolingual and Cross-Lingual Evaluation
Agirre, E., Banea, C., Cer, D. M., Diab, M., González-Agirre, A., Mihalcea, R. F., Rigau, G., & Wiebe, J. (2016) · International Workshop on Semantic Evaluation (SemEval) · doi:10.18653/v1/s16-1081 · Cited by 527 · see citing works
SemEval-2014 Task 10: Multilingual Semantic Textual Similarity
Agirre, E., Banea, C., Cardie, C., Cer, D. M., Diab, M., González-Agirre, A., Guo, W., Mihalcea, R. F., Rigau, G., & Wiebe, J. (2014) · International Workshop on Semantic Evaluation (SemEval) · doi:10.3115/v1/s14-2010 · Cited by 492 · see citing works
SemEval-2015 Task 2: Semantic Textual Similarity, English, Spanish and Pilot on Interpretability
Agirre, E., Banea, C., Cardie, C., Cer, D. M., Diab, M., González-Agirre, A., Guo, W., López-Gazpio, I., Maritxalar, M., Mihalcea, R. F., et al. (2015) · International Workshop on Semantic Evaluation (SemEval) · doi:10.18653/v1/s15-2045 · Cited by 488 · see citing works
SentEval: An Evaluation Toolkit for Universal Sentence Representations
Conneau, A., & Kiela, D. (2018) · arXiv · doi:10.48550/arxiv.1803.05449 · Cited by 477 · see citing works
SentEval: An Evaluation Toolkit for Universal Sentence Representations
Conneau, A., & Kiela, D. (2018) · Language Resources and Evaluation Conference (LREC) · doi:10.63317/2gscircifffd · Cited by 477 · see citing works
Representation Degeneration Problem in Training Natural Language Generation ModelsPreprint
Gao, J., He, D., Tan, X., Qin, T., Wang, L.-W., & Liu, T.-Y. (2019) · International Conference on Learning Representations · doi:10.48550/arxiv.1907.12009 · Cited by 437 · see citing works
*SEM 2013 shared task: Semantic Textual Similarity
Agirre, E., Cer, D. M., Diab, M., González-Agirre, A., & Guo, W. (2013) · Joint Conference on Lexical and Computational Semantics · Cited by 360 · see citing works
SemEval-2017 Task 1: Semantic Textual Similarity - Multilingual and Cross-lingual Focused Evaluation
Cer, D. M., Diab, M., Agirre, E., et al. (2017) · HAL (Le Centre pour la Communication Scientifique Directe) · doi:10.48550/arxiv.1708.00055 · Cited by 306 · see citing works
From Variational to Deterministic AutoencodersPreprint
Ghosh, P. S., Sajjadi, M. S. M., Vergari, A., Black, M. J., & Schölkopf, B. (2019) · arXiv (Cornell University) · doi:10.48550/arxiv.1903.12436 · Cited by 163 · see citing works
All-but-the-Top: Simple and Effective Postprocessing for Word RepresentationsPreprint
Mu, J., & Viswanath, P. (2018) · arXiv (Cornell University) · Cited by 128 · see citing works
All-but-the-Top: Simple and Effective Postprocessing for Word RepresentationsPreprint
Mu, J., Bhat, S., & Viswanath, P. (2017) · arXiv (Cornell University) · doi:10.48550/arxiv.1702.01417 · Cited by 126 · see citing works
Towards Understanding Linear Word Analogies
Ethayarajh, K., Duvenaud, D., & Hirst, G. (2019) · Proceedings of the conference - Association for Computational Linguistics. Meeting · doi:10.18653/v1/p19-1315 · Cited by 95 · see citing works
From Variational to Deterministic Autoencoders
Ghosh, P. S., Sajjadi, M. S. M., Vergari, A., Black, M. J., & Schölkopf, B. (2020) · International Conference on Learning Representations · Cited by 72 · see citing works
Taming VAEsPreprint
Rezende, D. J., & Viola, F. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1810.00597 · Cited by 60 · see citing works
A Surprisingly Effective Fix for Deep Latent Variable Modeling of Text
Li, B., He, J., Neubig, G., et al. (2019) · doi:10.18653/v1/d19-1370 · Cited by 59 · see citing works
Self-citationImproving Neural Language Generation with Spectrum Control
Wang, L., Huang, J., Huang, K., Hu, Z., Wang, G., & Gu, Q. (2020) · International Conference on Learning Representations · Cited by 53 · see citing works
Normalizing Flows: Introduction and Ideas.Preprint
Kobyzev, I., Prince, S., & Brubaker, M. A. (2019) · arXiv (Cornell University) · Cited by 34 · see citing works
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
Flagged first, then by citations
Research works printed but not linked: 5
- Alexis Conneau, Douwe Kiela, Holger Schwenk, Loïc Barrault, and Antoine Bordes. 2017. Supervised learning of universal sentence representations from natural language inference data. In Proceedings of EMNLP.
- Durk P Kingma and Prafulla Dhariwal. 2018. Glow: Generative flow with invertible 1x1 convolutions. In Proceedings of NeurIPS.
- Lingxiao Wang, Jing Huang, Kevin Huang, Ziniu Hu, Guangtao Wang, and Quanquan Gu. 2020. Improv- ing neural language generation with spectrum con- trol. In Proceedings of ICLR.
- Zhilin Yang, Zihang Dai, Ruslan Salakhutdinov, and William W Cohen. 2018. Breaking the softmax bot- tleneck: A high-rank rnn language model. In Pro- ceedings of ICLR.
- Zhilin Yang, Zihang Dai, Yiming Yang, Jaime Car- bonell, Russ R Salakhutdinov, and Quoc V Le. 2019. XLNet: Generalized autoregressive pretrain- ing for language understanding. In Proceedings of NeurIPS.
Cited by
Sorted by citations
Generating Datasets with Pretrained Language Models
Schick, T., & Schütze, H. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.555 · Cited by 122 · see citing works
Towards General Text Embeddings with Multi-stage Contrastive Learning
Li, Z., Zhang, X., Zhang, Y., Long, D., Xie, P., & Zhang, M. (2023) · arXiv · doi:10.48550/arxiv.2308.03281 · Cited by 79 · see citing works
Towards a holistic framework for multimodal LLM in 3D brain CT radiology report generation
Li, C.-Y., Chang, K.-J., Yang, C.-F., Wu, H.-Y., Chen, W., Bansal, H., Chen, L., Yang, Y.-P., Chen, Y.-C., Chen, S.-P., et al. (2025) · Nature Communications · doi:10.1038/s41467-025-57426-0 · Cited by 75 · see citing works
Self-citationContrastive Learning Models for Sentence Representations
Xu, L., Xie, H., Li, Z., Wang, F. L., Wang, W., & Li, Q. (2023) · ACM Transactions on Intelligent Systems and Technology · doi:10.1145/3593590 · Cited by 50 · see citing works
Li, L., Kong, S., Zhao, H., Li, C., Teng, Y., & Wang, Y. (2025) · Psychiatry and Clinical Neurosciences · doi:10.1111/pcn.13781 · Cited by 10 · see citing works
Self-citation
- Authors
- Li, B., Zhou, H., He, J., Wang, M., Yang, Y., & Li, L.
- Affiliations
- Carnegie Mellon University (US)
- Corresponding author
- Yang, Yiming
- Venue
- Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)
- Published
- 2020
- Pages
- 9119-9130
- OpenAlex
- W3105816068
- Cited by
- 593
- References
- 36
- Citations by year
- 2026: 46 · 2025: 70 · 2024: 92 · 2023: 182 · 2022: 128 · 2021: 73 · 2020: 2
- FWCI
- 36.76
- 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)
- Rare & Special e-Zone (The Hong Kong University of Science and Technology) (submittedVersion)
- arXiv (Cornell University)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Physical Sciences › Computer Science › Artificial Intelligence › Sentiment Analysis and Opinion Mining
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Multimodal Machine Learning Applications
- Keywords
- sentence embeddings, BERT, semantic similarity, normalizing flows, unsupervised learning, semantic textual similarity
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
