Frequency Effects on Syntactic Rule Learning in Transformers
Open accessThe DOI matches its Crossref record; 13 of the 13 references it cites were found in a registry or an index.
Abstract
Pre-trained language models perform well on a variety of linguistic tasks that require symbolic reasoning, raising the question of whether such models implicitly represent abstract symbols and rules.We investigate this question using the case study of BERT's performance on English subject-verb agreement.Unlike prior work, we train multiple instances of BERT from scratch, allowing us to perform a series of controlled interventions at pre-training time.We show that BERT often generalizes well to subject-verb pairs that never occurred in training, suggesting a degree of rule-governed behavior.We also find, however, that performance is heavily influenced by word frequency, with experiments showing that both the absolute frequency of a verb form, as well as the frequency relative to the alternate inflection, are causally implicated in the predictions BERT makes at inference time.Closer analysis of these frequency effects reveals that BERT's behavior is consistent with a system that correctly applies the SVA rule in general but struggles to overcome strong training priors and to estimate agreement features (singular vs. plural) on infrequent lexical items.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
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. (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
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
Connectionism and cognitive architecture: A critical analysis
Fodor, J. A., & Pylyshyn, Z. W. (1988) · Cognition · doi:10.1016/0010-0277(88)90031-5 · Cited by 4,022
On the proper treatment of connectionism
Smolensky, P. (1988) · Behavioral and Brain Sciences · doi:10.1017/s0140525x00052432 · Cited by 2,441
Zhu, Y., Kiros, R., Zemel, R., Salakhutdinov, R., Urtasun, R., Torralba, A., & Fidler, S. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.11 · Cited by 2,264 · cited 3×
“The original BERT model was trained on both English Wikipedia and BooksCorpus (Zhu et al., 2015).”
MethodInfluentialA 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 · cited 2×
“While previous studies have attempted to address such questions, particularly in relation to BERT (Rogers et al., 2020), prior work has generally not analyzed the relationship between the model’s pre-training data and its behavior.”
ResultThe ubiquity of frequency effects in first language acquisition
AMBRIDGE, B., KIDD, E., ROWLAND, C. F., & THEAKSTON, A. L. (2015) · Journal of Child Language · doi:10.1017/s030500091400049x · Cited by 589 · cited 3×
“High-frequency forms cause errors when a competing, lower-frequency form is the target (Ambridge et al., 2015).”
BackgroundInfluentialAssessing BERT's Syntactic Abilities
Goldberg, Y. (2019) · arXiv · doi:10.48550/arxiv.1901.05287 · Cited by 385 · cited 7×
“Following the setup from Goldberg (2019), each test instance consists of a sentence in which a verb has been masked out, and BERT’s masked language modeling (MLM) parameters are used to score whether the singular or plural form of the verb is a better fit for the masked position.”
BackgroundMethodInfluentialTargeted Syntactic Evaluation of Language Models
Marvin, R., & Linzen, T. (2018) · Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/d18-1151 · Cited by 334 · cited 2×
“The success of contemporary neural language models such as BERT (Devlin et al., 2019) on language understanding tasks, as well as in more targeted linguistic evaluations (Marvin and Linzen, 2018; Goldberg, 2019), raises the question of whether these systems acquire such symbolic rules.”
BackgroundMethodSelf-citationSyntactic Structure from Deep Learning
Linzen, T., & Baroni, M. (2021) · Annual Review of Linguistics · doi:10.1146/annurev-linguistics-032020-051035 · Cited by 202
Self-citation
Cited by
Emergent Abilities of Large Language Models
Wei, J., Tay, Y., Bommasani, R., Raffel, C., Zoph, B., Borgeaud, S., Yogatama, D., Bosma, M., Zhou, D., Metzler, D., et al. (2022) · arXiv · doi:10.48550/arxiv.2206.07682 · Cited by 1,264
“Overall, more work is needed to tease apart what enables scale to unlock emergent abilities.”
Self-citationLanguage Model Behavior: A Comprehensive Survey
Chang, T. A., & Bergen, B. K. (2024) · Computational Linguistics · doi:10.1162/coli_a_00492 · Cited by 105
“(Wei et al. 2021), with further degradations if there is an intervening clause between the subject and verb as in Example 4 (Lasri, Lenci, and Poibeau 2022a).”
BackgroundImpact of Pretraining Term Frequencies on Few-Shot Numerical Reasoning
Razeghi, Y., Logan IV, R. L., Gardner, M., & Singh, S. (2022) · Findings of the Association for Computational Linguistics: EMNLP 2022 · doi:10.18653/v1/2022.findings-emnlp.59 · Cited by 89
“Kassner et al. (2020) and Wei et al. (2021) perform controlled experiments varying pretraining data to characterize the extent to which pretraining affects LMs’ ability to memorize and reason with facts as well as learn generalizable syntax rules.”
BackgroundMcCoy, R. T., Smolensky, P., Linzen, T., Gao, J., & Celikyilmaz, A. (2023) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00567 · Cited by 70
“Though many of the abstractions that we study have been discussed in prior analyses of LMs, the only one for which prior work has enforced novelty is subject-verb agreement (Wei et al., 2021).”
BackgroundSelf-citationMcCoy, R. T., Smolensky, P., Linzen, T., Gao, J., & Celikyilmaz, A. (2021) · arXiv · doi:10.48550/arxiv.2111.09509 · Cited by 70
Self-citation
Metadata
- Authors
- Wei, J., Garrette, D., Linzen, T., & Pavlick, E.
- Venue
- Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing
- Published
- 2021
- Pages
- 932-948
- OpenAlex
- W3200809495
- Cited by
- 42
- References
- 13
- FWCI
- 4.07
- Type
- conference paper
- Language
- en
- Open access
- hybrid
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
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Frequency Effects on Syntactic Rule Learning in Transformers. https://app.cochat.ai/papers/10.18653/v1/2021.emnlp-main.72
- The paper
- Wei, J., Garrette, D., Linzen, T., & Pavlick, E. (2021). Frequency Effects on Syntactic Rule Learning in Transformers. Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing. https://doi.org/10.18653/v1/2021.emnlp-main.72
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.