The DOI matches its Crossref record; 51 of the 51 references it cites were found in a registry or an index.
Deduplicating Training Data Makes Language Models Better
Open accessAbstract
Katherine Lee, Daphne Ippolito, Andrew Nystrom, Chiyuan Zhang, Douglas Eck, Chris Callison-Burch, Nicholas Carlini. Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2022.
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
“Data deduplication has become a key tool for language model projects following research showing that it both improves performance on downstream tasks (Lee et al., 2022; Zhang et al., 2021) and decreases memorization of training data (Kandpal et al., 2022).”
From The BigScience ROOTS Corpus: A 1.6TB Composite Multilingual Dataset (2023) · cites this paper as Lee et al., 2022
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“To ensure reliable evaluation and improve performance [26], we deduplicate documents by title and utilize the predefined training and validation splits.”
From Lightweight and Direct Document Relevance Optimization for Generative Information Retrieval (2025) · cites this paper as [26]
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“For example, the probability of reproducing training data has been shown to correlate with its frequency of appearance [12].”
From Hallucinations in Bibliographic Recommendation: Citation Frequency as a Proxy for Training Data Redundancy (2025) · cites this paper as [18]
The cited source says what the claim attributes to it.
In the citing paper
“Deduplication is a powerful strategy for improving dataset quality (Li et al., 2024; Lee et al., 2022; Penedo et al., 2024b; Fang et al., 2025; Liu et al., 2024).”
From OpenThoughts: Data Recipes for Reasoning Models (2025) · cites this paper as Lee et al., 2022
The check wasn’t confident either way, so it gives no verdict — never disputed.
In the citing paper
“Moreover, it is provided only on the complete text generated by the model (trajectory) and not on specific parts of the text (particular state) [92,138].”
From RLHF Deciphered: A Critical Analysis of Reinforcement Learning from Human Feedback for LLMs (2026) · cites this paper as [92]
Experimental — accuracy not yet measured
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
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
Language Models are Few-Shot Learners
Brown, T. B., Mann, B., Ryder, N., Subbiah, M., Kaplan, J., Dhariwal, P., Neelakantan, A., Shyam, P., Sastry, G., Askell, A., et al. (2020) · arXiv · doi:10.48550/arxiv.2005.14165 · Cited by 23,372
Space/time trade-offs in hash coding with allowable errors
Bloom, B. (1970) · CACM · doi:10.1145/362686.362692 · Cited by 8,424
On the Dangers of Stochastic Parrots: Can Language Models Be Too Big? 🦜
Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021) · Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency · doi:10.1145/3442188.3445922 · Cited by 6,988 · cited 4×
“And so because these datasets are up to 19% smaller, even including the deduplication runtime itself, training on dedupli-cated datasets directly reduces the training cost in terms of time, dollar, and the environment (Bender et al., 2021; Strubell et al., 2019; Patterson et al., 2021).”
BackgroundInfluentialMembership Inference Attacks Against Machine Learning Models
Shokri, R., Stronati, M., Song, C., & Shmatikov, V. (2017) · 2017 IEEE Symposium on Security and Privacy (SP) · doi:10.1109/sp.2017.41 · Cited by 4,754
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
Energy and Policy Considerations for Deep Learning in NLP
Strubell, E., Ganesh, A., & McCallum, A. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/p19-1355 · Cited by 3,684
Transformer-XL: Attentive language models beyond a fixed-length context
Dai, Z., Yang, Z., Yang, Y., Carbonell, J., Le, Q., & Salakhutdinov, R. (2019) · Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/p19-1285 · Cited by 3,169
Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., Daumé, H., & Crawford, K. (2018) · arXiv · doi:10.48550/arxiv.1803.09010 · Cited by 2,474
Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., III, H. D., & Crawford, K. (2021) · Communications of the ACM · doi:10.1145/3458723 · Cited by 2,469
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
Suffix Arrays: A New Method for On-Line String Searches
Manber, U., & Myers, E. W. (1993) · SIAM Journal on Computing · doi:10.1137/0222058 · Cited by 1,915
Linear pattern matching algorithms
Weiner, P. D. (1973) · Annual Symposium on Switching & Automata Theory · doi:10.1109/swat.1973.13 · Cited by 1,829
On the resemblance and containment of documents
Broder, A. (2002) · Compression and Complexity of Sequences (SEQUENCES) · doi:10.1109/sequen.1997.666900 · Cited by 1,818
mT5: A Massively Multilingual Pre-trained Text-to-Text Transformer
Xue, L., Constant, N., Roberts, A., Kale, M., Al-Rfou, R., Siddhant, A., Barua, A., & Raffel, C. (2021) · Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies · doi:10.18653/v1/2021.naacl-main.41 · Cited by 1,629
Transformer-XL: Attentive Language Models Beyond a Fixed-Length ContextPreprint
Dai, Z., Yang, Z., Yang, Y., Carbonell, J., Le, Q. V., & Salakhutdinov, R. (2019) · arXiv (Cornell University) · doi:10.4230/lipics.ecrts.2023.7 · Cited by 1,476
Adafactor: Adaptive Learning Rates with Sublinear Memory CostPreprint
Shazeer, N., & Stern, M. (2018) · International Conference on Machine Learning · doi:10.48550/arxiv.1804.04235 · Cited by 1,410
mT5: A massively multilingual pre-trained text-to-text transformerPreprint
Xue, L., Constant, N., Roberts, A., Kale, M., Al‐Rfou, R., Siddhant, A., Barua, A., & Raffel, C. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2010.11934 · Cited by 1,324
A Closer Look at Memorization in Deep Networks
Arpit, D., Jastrzębski, S., Ballas, N., Krueger, D., Bengio, E., Kanwal, M. S., Maharaj, T., Fischer, A., Courville, A., Bengio, Y., et al. (2017) · arXiv · doi:10.48550/arxiv.1706.05394 · Cited by 1,307
One Billion Word Benchmark for Measuring Progress in Statistical Language Modeling
Chelba, C., Mikolov, T., Schuster, M., Ge, Q., Brants, T., Koehn, P., & Robinson, T. (2013) · arXiv · doi:10.48550/arxiv.1312.3005 · Cited by 1,051
Bender, E. M., & Friedman, B. (2018) · Transactions of the Association for Computational Linguistics · doi:10.1162/tacl_a_00041 · Cited by 854
Extracting Training Data from Large Language Models
Carlini, N., Tramer, F., Wallace, E., Jagielski, M., Herbert-Voss, A., Lee, K., Roberts, A., Brown, T., Song, D., Erlingsson, U., et al. (2020) · arXiv · doi:10.48550/arxiv.2012.07805 · Cited by 727
Self-citationUniversal Adversarial Triggers for Attacking and Analyzing NLP
Wallace, E., Feng, S., Kandpal, N., Gardner, M., & Singh, S. (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/d19-1221 · Cited by 642
One billion word benchmark for measuring progress in statistical language modeling
Chelba, C., Mikolov, T., Schuster, M., Ge, Q., Brants, T., Koehn, P., & Robinson, T. (2014) · Interspeech 2014 · doi:10.21437/interspeech.2014-564 · Cited by 627
Defending Against Neural Fake News
Zellers, R., Holtzman, A., Rashkin, H., Bisk, Y., Farhadi, A., Roesner, F., & Choi, Y. (2019) · arXiv · doi:10.48550/arxiv.1905.12616 · Cited by 569
The Pile: An 800GB Dataset of Diverse Text for Language Modeling
Gao, L., Biderman, S., Black, S., Golding, L., Hoppe, T., Foster, C., Phang, J., He, H., Thite, A., Nabeshima, N., et al. (2021) · arXiv · doi:10.48550/arxiv.2101.00027 · Cited by 492
Carbon Emissions and Large Neural Network Training
Patterson, D., Gonzalez, J., Le, Q., Liang, C., Munguia, L.-M., Rothchild, D., So, D., Texier, M., & Dean, J. (2021) · arXiv · doi:10.48550/arxiv.2104.10350 · Cited by 486
Simple Linear Work Suffix Array Construction
Kärkkäinen, J., & Sanders, P. W. (2003) · Lecture notes in computer science · doi:10.1007/3-540-45061-0_73 · Cited by 404
GPT-Neo: Large Scale Autoregressive Language Modeling with Mesh-Tensorflow
Black, S., Leo, G., Wang, P., Leahy, C., & Biderman, S. (2021) · Zenodo · doi:10.5281/zenodo.5297715 · Cited by 324
A Simple Method for Commonsense Reasoning
Trinh, T. H., & Le, Q. V. (2018) · arXiv · doi:10.48550/arxiv.1806.02847 · Cited by 293
The adverse effects of code duplication in machine learning models of code
Allamanis, M. (2019) · Proceedings of the 2019 ACM SIGPLAN International Symposium on New Ideas, New Paradigms, and Reflections on Programming and Software · doi:10.1145/3359591.3359735 · Cited by 284
Space Efficient Linear Time Construction of Suffix Arrays
Ko, P., & Aluru, S. (2003) · Lecture notes in computer science · doi:10.1007/3-540-44888-8_15 · Cited by 225
Using Suffix Arrays to Compute Term Frequency and Document Frequency for All Substrings in a Corpus
Yamamoto, M., & Church, K. (2001) · Computational Linguistics · doi:10.1162/089120101300346787 · Cited by 177
Zeng, W., Ren, X., Su, T., Wang, H., Liao, Y., Wang, Z., Jiang, X., Yang, Z., Wang, K., Zhang, X., et al. (2021) · arXiv · doi:10.48550/arxiv.2104.12369 · Cited by 172
Self-citationWhat Does it Mean for a Language Model to Preserve Privacy?
Brown, H., Lee, K., Mireshghallah, F., Shokri, R., & Tramèr, F. (2022) · 2022 ACM Conference on Fairness Accountability and Transparency · doi:10.1145/3531146.3534642 · Cited by 165
Self-citationWhat Neural Networks Memorize and Why: Discovering the Long Tail via Influence EstimationPreprint
Feldman, V., & Zhang, C. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2008.03703 · Cited by 146
Self-citationLinear Suffix Array Construction by Almost Pure Induced-Sorting
Nong, G., Zhang, S., & Chan, W. H. (2009) · DCC · doi:10.1109/dcc.2009.42 · Cited by 143
Towards Controllable Biases in Language Generation
Sheng, E., Chang, K.-W., Natarajan, P., & Peng, N. (2020) · Findings of the Association for Computational Linguistics: EMNLP (Findings of EMNLP) · doi:10.18653/v1/2020.findings-emnlp.291 · Cited by 122
A new suffix tree similarity measure for document clustering
Chim, H., & Deng, X. (2007) · International World Wide Web Conference (WWW) · doi:10.1145/1242572.1242590 · Cited by 108
Detecting Overfitting of Deep Generative Networks via Latent Recovery
Webster, R., Rabin, J., Simon, L., & Jurie, F. (2019) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2019.01153 · Cited by 101
On the geometry of generalization and memorization in deep neural networksPreprint
Stephenson, C., Padhy, S., Ganesh, A., Hui, Y., Tang, H., & Chung, S. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2105.14602 · Cited by 58
Bandy, J., & Vincent, N. (2021) · arXiv · doi:10.48550/arxiv.2105.05241 · Cited by 36
Identifying and characterizing highly similar notes in big clinical note datasets
Gabriel, R. A., Kuo, T.-T., McAuley, J., & Hsu, C. (2018) · Journal of Biomedical Informatics · doi:10.1016/j.jbi.2018.04.009 · Cited by 34
Wiki-40B: Multilingual Language Model Dataset
Guo, M., Dai, Z., Vrandečić, D., & Al‐Rfou, R. (2020) · Language Resources and Evaluation · Cited by 33
(2021) · arXiv (Cornell University) · Cited by 21
Connected Components at Scale via Local ContractionsPreprint
Łącki, J., Mirrokni, V., & Włodarczyk, M. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1807.10727 · Cited by 19
Towards Controllable Biases in Language GenerationPreprint
Sheng, E., Chang, K.-W., Natarajan, P., et al. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2005.00268 · Cited by 11
Deduplication of Scholarly Documents using Locality Sensitive Hashing and Word Embeddings
Gyawali, B., Anastasiou, L., & Knoth, P. (2020) · Open Research Online (The Open University) · Cited by 10
Understanding Invariance via Feedforward Inversion of Discriminatively Trained ClassifiersPreprint
Teterwak, P., Zhang, C., Krishnan, D., & Mozer, M. C. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2103.07470 · Cited by 4
Self-citationDocumenting Large Webtext Corpora: A Case Study on the Colossal Clean Crawled Corpus
Dodge, J., Sap, M., Marasović, A., et al. (2021) · arXiv · doi:10.48550/arxiv.2104.08758
Universal Adversarial Triggers for Attacking and Analyzing NLP
Wallace, E., Feng, S., Kandpal, N., Gardner, M., & Singh, S. (2019) · arXiv · doi:10.48550/arxiv.1908.07125
Flagged first, then by citations
Research works printed but not linked: 7
- PMLR.
- David Graff, Junbo Kong, Ke Chen, and Kazuaki Maeda. 2003. English gigaword. Linguistic Data Consortium, Philadelphia, 4(1):34.
- Paul Jaccard. 1912. The distribution of the flora in the alpine zone. New phytologist, 11(2):37-50. 10 8433
- Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language models are unsupervised multitask learners. OpenAI blog, 1(8):9.
- Colin Raffel, Noam Shazeer, Adam Roberts, Kather- ine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J. Liu. 2020. Exploring the limits of transfer learning with a unified text-to- text transformer. Journal of Machine Learning Re- search, 21(140):1-67.
- PMLR.
- Yannick Versley and Yana Panchenko. 2012. Not just bigger: Towards better-quality web corpora. In Pro- ceedings of the seventh Web as Corpus Workshop (WAC7), pages 44-52.
Cited by
Sorted by citations
Survey of Hallucination in Natural Language Generation
Ji, Z., Lee, N., Frieske, R., Yu, T., Su, D., Xu, Y., Ishii, E., Bang, Y. J., Madotto, A., & Fung, P. (2023) · ACM Computing Surveys · doi:10.1145/3571730 · Cited by 5,025
“[96] show that duplicated examples from the pretraining corpus bias the model to favor generating repeats of the memorized phrases from the duplicated examples.”
BackgroundOn the Opportunities and Risks of Foundation Models
Bommasani, R., Hudson, D. A., Adeli, E., Altman, R., Arora, S., von Arx, S., Bernstein, M. S., Bohg, J., Bosselut, A., Brunskill, E., et al. (2021) · arXiv · doi:10.48550/arxiv.2108.07258 · Cited by 2,676
Huang, L., Yu, W., Ma, W., Zhong, W., Feng, Z., Wang, H., Chen, Q., Peng, W., Feng, X., Qin, B., et al. (2025) · ACM Transactions on Information Systems · doi:10.1145/3703155 · Cited by 4,688
“However, the inherent memo-rization capability becomes problematic in the context of duplicated information present within pre-training data (Lee et al., 2022a; Kandpal et al., 2023; Paullada et al., 2021).”
A survey on large language model (LLM) security and privacy: The Good, The Bad, and The Ugly
Yao, Y., Duan, J., Xu, K., Cai, Y., Sun, Z., & Zhang, Y. (2024) · High-Confidence Computing · doi:10.1016/j.hcc.2024.100211 · Cited by 1,083
Multitask Prompted Training Enables Zero-Shot Task Generalization
Sanh, V., Webson, A., Raffel, C., Bach, S. H., Sutawika, L., Alyafeai, Z., Chaffin, A., Stiegler, A., Scao, T. L., Raja, A., et al. (2021) · arXiv · doi:10.48550/arxiv.2110.08207 · Cited by 612
Metadata
- Authors
- Lee, K., Ippolito, D., Nystrom, A., Zhang, C., Eck, D., Callison-Burch, C., & Carlini, N.
- Affiliations
- Google (United States) (US)
- University of Pennsylvania (US)
- Corresponding author
- Ippolito, Daphne
- Venue
- Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Published
- 2022
- Pages
- 8424-8445
- ISSN
- 0736-587X (ISSN-L)
- OpenAlex
- W3177765786
- Cited by
- 304
- References
- 50
- Citations by year
- 2026: 62 · 2025: 79 · 2024: 65 · 2023: 64 · 2022: 27
- FWCI
- 32.49
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Proceedings of the conference - Association for Computational Linguistics. Meeting (publishedVersion, cc-by)
- 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 › Machine Learning and Data Classification
- Keywords
- language model training, training data quality, duplicate removal
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
