Analyzing Transformers in Embedding Space
Open accessThe DOI matches its Crossref record; 20 of the 20 references it cites were found in a registry or an index.
Abstract
Understanding Transformer-based models has attracted significant attention, as they lie at the heart of recent technological advances across machine learning.While most interpretability methods rely on running models over inputs, recent work has shown that an inputindependent approach, where parameters are interpreted directly without a forward/backward pass is feasible for some Transformer parameters, and for two-layer attention networks.In this work, we present a conceptual framework where all parameters of a trained Transformer are interpreted by projecting them into the embedding space, that is, the space of vocabulary items they operate on.Focusing mostly on GPT-2 for this paper, we provide diverse evidence to support our argument.First, an empirical analysis showing that parameters of both pretrained and fine-tuned models can be interpreted in embedding space.Second, we present two applications of our framework: (a) aligning the parameters of different models that share a vocabulary, and (b) constructing a classifier without training by "translating" the parameters of a fine-tuned classifier to parameters of a different model that was only pretrained.Overall, our findings show that at least in part, we can abstract away model specifics and understand Transformers in the embedding space.EA EB Layer 18 Head 1 ('women', ' Marie') (' actresses', ' Marie') ('women', ' Anne') ('Women', ' Anne') ('woman', ' Marie') ('Women', ' Marie')
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. (2018) · arXiv · doi:10.48550/arxiv.1810.04805 · Cited by 73,173
An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale
Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., et al. (2020) · arXiv · doi:10.48550/arxiv.2010.11929 · Cited by 47,472
Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer
Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., Zhou, Y., Li, W., & Liu, P. J. (2019) · arXiv · doi:10.48550/arxiv.1910.10683 · Cited by 10,929
Transformers: State-of-the-Art Natural Language Processing
Wolf, T., Debut, L., Sanh, V., Chaumond, J., Delangue, C., Moi, A., Cistac, P., Rault, T., Louf, R., Funtowicz, M., et al. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations · doi:10.18653/v1/2020.emnlp-demos.6 · Cited by 8,741
THE DISTRIBUTION OF THE FLORA IN THE ALPINE ZONE. 1
Jaccard, P. (1912) · New Phytologist · doi:10.1111/j.1469-8137.1912.tb05611.x · Cited by 4,561
wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations
Baevski, A., Zhou, H., Mohamed, A., & Auli, M. (2020) · arXiv · doi:10.48550/arxiv.2006.11477 · Cited by 2,369
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
BERT Rediscovers the Classical NLP Pipeline
Tenney, I., Das, D., & Pavlick, E. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/p19-1452 · Cited by 1,307
Analyzing Multi-Head Self-Attention: Specialized Heads Do the Heavy Lifting, the Rest Can Be Pruned
Voita, E., Talbot, D., Moiseev, F., Sennrich, R., & Titov, I. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/p19-1580 · Cited by 1,146
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
Understanding image representations by measuring their equivariance and equivalence
Lenc, K., & Vedaldi, A. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298701 · Cited by 366
Transformer Feed-Forward Layers Are Key-Value Memories
Geva, M., Schuster, R., Berant, J., & Levy, O. (2020) · arXiv · doi:10.48550/arxiv.2012.14913 · Cited by 353
Self-citationTransformer Feed-Forward Layers Are Key-Value Memories
Geva, M., Schuster, R., Berant, J., & Levy, O. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.446 · Cited by 353
Self-citationKnowledge Neurons in Pretrained Transformers
Dai, D., Dong, L., Hao, Y., Sui, Z., Chang, B., & Wei, F. (2022) · Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2022.acl-long.581 · Cited by 187
Transformer Feed-Forward Layers Build Predictions by Promoting Concepts in the Vocabulary Space
Geva, M., Caciularu, A., Wang, K. R., & Goldberg, Y. (2022) · arXiv · doi:10.48550/arxiv.2203.14680 · Cited by 123
Self-citationAugmenting Self-attention with Persistent Memory
Sukhbaatar, S., Grave, E., Lample, G., Jegou, H., & Joulin, A. (2019) · arXiv · doi:10.48550/arxiv.1907.01470 · Cited by 75
Knowledge Neurons in Pretrained Transformers
Dai, D., Dong, L., Hao, Y., Sui, Z., Chang, B., & Wei, F. (2021) · arXiv · doi:10.48550/arxiv.2104.08696 · Cited by 39
The MultiBERTs: BERT Reproductions for Robustness Analysis
Sellam, T., Yadlowsky, S., Wei, J., Saphra, N., D'Amour, A., Linzen, T., Bastings, J., Turc, I., Eisenstein, J., Das, D., et al. (2021) · arXiv · doi:10.48550/arxiv.2106.16163 · Cited by 37
Revisiting Model Stitching to Compare Neural Representations
Bansal, Y., Nakkiran, P., & Barak, B. (2021) · arXiv · doi:10.48550/arxiv.2106.07682 · Cited by 26
Voita, E., Sennrich, R., & Titov, I. (2019) · arXiv · doi:10.48550/arxiv.1909.01380 · Cited by 17
Cited by
(2024) · doi:10.1109/access.2024.3365742 · Cited by 934
Open Problems in Mechanistic Interpretability
Sharkey, L., Chughtai, B., Batson, J., Lindsey, J., Wu, J., Bushnaq, L., Goldowsky-Dill, N., Heimersheim, S., Ortega, A., Bloom, J., et al. (2025) · arXiv · doi:10.48550/arxiv.2501.16496 · Cited by 291
Self-citationThe Hydra Effect: Emergent Self-repair in Language Model Computations
McGrath, T., Rahtz, M., Kramar, J., Mikulik, V., & Legg, S. (2023) · arXiv · doi:10.48550/arxiv.2307.15771 · Cited by 140
Language Model Behavior: A Comprehensive Survey
Chang, T. A., & Bergen, B. K. (2024) · Computational Linguistics · doi:10.1162/coli_a_00492 · Cited by 105
Jump to Conclusions: Short-Cutting Transformers with Linear Transformations
Yom Din, A., Karidi, T., Choshen, L., & Geva, M. (2024) · Proceedings of the Language Resources and Evaluation Conference · doi:10.63317/5griomztwyxs · Cited by 61
Self-citation
Metadata
- Authors
- Dar, G., Geva, M., Gupta, A., & Berant, J.
- Venue
- Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Published
- 2023
- Pages
- 16124-16170
- OpenAlex
- W4385571791
- Cited by
- 49
- References
- 20
- FWCI
- 9.26
- 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
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.