# Whose Language Counts as High Quality? Measuring Language Ideologies in Text Data Selection

arXiv: https://arxiv.org/abs/2201.10474 · HTML: https://mlsanity.com/paper/2201.10474/

## Key facts

- Cited by 53 papers in the map (data through 14 Oct 2026).
- Builds on 28 of its 98 references that are in the map.
- First posted on arXiv 25 Jan 2022.

## Builds on (most cited first, 28 of 28)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://mlsanity.com/paper/1910.13461/) (2019)
- [Unsupervised Cross-lingual Representation Learning at Scale](https://mlsanity.com/paper/1911.02116/) (2019)
- [Switch Transformers: Scaling to Trillion Parameter Models with Simple and Efficient Sparsity](https://mlsanity.com/paper/2101.03961/) (2021)
- [XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://mlsanity.com/paper/1906.08237/) (2019)
- [ALBERT: A Lite BERT for Self-supervised Learning of Language Representations](https://mlsanity.com/paper/1909.11942/) (2019)
- [The Pile: An 800GB Dataset of Diverse Text for Language Modeling](https://mlsanity.com/paper/2101.00027/) (2020)
- [Deep contextualized word representations](https://mlsanity.com/paper/1802.05365/) (2018)
- [Datasheets for Datasets](https://mlsanity.com/paper/1803.09010/) (2018)
- [RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models](https://mlsanity.com/paper/2009.11462/) (2020)
- [Scaling Language Models: Methods, Analysis & Insights from Training Gopher](https://mlsanity.com/paper/2112.11446/) (2021)
- [Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books](https://mlsanity.com/paper/1506.06724/) (2015)
- [Language (Technology) is Power: A Critical Survey of "Bias" in NLP](https://mlsanity.com/paper/2005.14050/) (2020)
- [CCNet: Extracting High Quality Monolingual Datasets from Web Crawl Data](https://mlsanity.com/paper/1911.00359/) (2019)
- [Persistent Anti-Muslim Bias in Large Language Models](https://mlsanity.com/paper/2101.05783/) (2021)
- [Annotators with Attitudes: How Annotator Beliefs And Identities Bias Toxic Language Detection](https://mlsanity.com/paper/2111.07997/) (2021)
- [Demographic Dialectal Variation in Social Media: A Case Study of African-American English](https://mlsanity.com/paper/1608.08868/) (2016)
- [Lessons from Archives: Strategies for Collecting Sociocultural Data in Machine Learning](https://mlsanity.com/paper/1912.10389/) (2019)
- [The Radicalization Risks of GPT-3 and Advanced Neural Language Models](https://mlsanity.com/paper/2009.06807/) (2020)
- [Predicting Factuality of Reporting and Bias of News Media Sources](https://mlsanity.com/paper/1810.01765/) (2018)
- [HTLM: Hyper-Text Pre-Training and Prompting of Language Models](https://mlsanity.com/paper/2107.06955/) (2021)
- [It's a Man's Wikipedia? Assessing Gender Inequality in an Online Encyclopedia](https://mlsanity.com/paper/1501.06307/) (2015)
- [Characterizing English Variation across Social Media Communities with BERT](https://mlsanity.com/paper/2102.06820/) (2021)
- [First Women, Second Sex: Gender Bias in Wikipedia](https://mlsanity.com/paper/1502.02341/) (2015)
- [An Empirical Exploration in Quality Filtering of Text Data](https://mlsanity.com/paper/2109.00698/) (2021)

## Built on by (most cited first, 30 of 53)

- [Language Model Tokenizers Introduce Unfairness Between Languages](https://mlsanity.com/paper/2305.15425/) (2023)
- [Do All Languages Cost the Same? Tokenization in the Era of Commercial Language Models](https://mlsanity.com/paper/2305.13707/) (2023)
- [A Pretrainer's Guide to Training Data: Measuring the Effects of Data Age, Domain Coverage, Quality, & Toxicity](https://mlsanity.com/paper/2305.13169/) (2023)
- [Rephrasing the Web: A Recipe for Compute and Data-Efficient Language Modeling](https://mlsanity.com/paper/2401.16380/) (2024)
- [Prompting PaLM for Translation: Assessing Strategies and Performance](https://mlsanity.com/paper/2211.09102/) (2022)
- [The PRISM Alignment Dataset: What Participatory, Representative and Individualised Human Feedback Reveals About the Subjective and Multicultural Alignment of Large Language Models](https://mlsanity.com/paper/2404.16019/) (2024)
- [Speak, Memory: An Archaeology of Books Known to ChatGPT/GPT-4](https://mlsanity.com/paper/2305.00118/) (2023)
- [Pile of Law: Learning Responsible Data Filtering from the Law and a 256GB Open-Source Legal Dataset](https://mlsanity.com/paper/2207.00220/) (2022)
- [NLPositionality: Characterizing Design Biases of Datasets and Models](https://mlsanity.com/paper/2306.01943/) (2023)
- [Language Contamination Helps Explain the Cross-lingual Capabilities of English Pretrained Models](https://mlsanity.com/paper/2204.08110/) (2022)
- [Re-contextualizing Fairness in NLP: The Case of India](https://mlsanity.com/paper/2209.12226/) (2022)
- [Multi-VALUE: A Framework for Cross-Dialectal English NLP](https://mlsanity.com/paper/2212.08011/) (2022)
- [Dated Data: Tracing Knowledge Cutoffs in Large Language Models](https://mlsanity.com/paper/2403.12958/) (2024)
- [BigBIO: A Framework for Data-Centric Biomedical Natural Language Processing](https://mlsanity.com/paper/2206.15076/) (2022)
- [Ecosystem Graphs: The Social Footprint of Foundation Models](https://mlsanity.com/paper/2303.15772/) (2023)
- [On the application of Large Language Models for language teaching and assessment technology](https://mlsanity.com/paper/2307.08393/) (2023)
- [AboutMe: Using Self-Descriptions in Webpages to Document the Effects of English Pretraining Data Filters](https://mlsanity.com/paper/2401.06408/) (2024)
- [Position: Key Claims in LLM Research Have a Long Tail of Footnotes](https://mlsanity.com/paper/2308.07120/) (2023)
- [Measuring Data](https://mlsanity.com/paper/2212.05129/) (2022)
- [Data Management For Training Large Language Models: A Survey](https://mlsanity.com/paper/2312.01700/) (2023)
- [ChineseWebText: Large-scale High-quality Chinese Web Text Extracted with Effective Evaluation Model](https://mlsanity.com/paper/2311.01149/) (2023)
- [Persona-Augmented Benchmarking: Evaluating LLMs Across Diverse Writing Styles](https://mlsanity.com/paper/2507.22168/) (2025)
- [TADA: Task-Agnostic Dialect Adapters for English](https://mlsanity.com/paper/2305.16651/) (2023)
- [Balancing Enhancement, Harmlessness, and General Capabilities: Enhancing Conversational LLMs with Direct RLHF](https://mlsanity.com/paper/2403.02513/) (2024)
- [Who's in and who's out? A case study of multimodal CLIP-filtering in DataComp](https://mlsanity.com/paper/2405.08209/) (2024)
- [Data, Data Everywhere: A Guide for Pretraining Dataset Construction](https://mlsanity.com/paper/2407.06380/) (2024)
- [Beyond the Binary: Capturing Diverse Preferences With Reward Regularization](https://mlsanity.com/paper/2412.03822/) (2024)
- [Task-Agnostic Low-Rank Adapters for Unseen English Dialects](https://mlsanity.com/paper/2311.00915/) (2023)
- [Evaluating Task-oriented Dialogue Systems: A Systematic Review of Measures, Constructs and their Operationalisations](https://mlsanity.com/paper/2312.13871/) (2023)
- [Natural Language Processing RELIES on Linguistics](https://mlsanity.com/paper/2405.05966/) (2024)

## Abstract (arXiv)

Language models increasingly rely on massive web dumps for diverse text data. However, these sources are rife with undesirable content. As such, resources like Wikipedia, books, and newswire often serve as anchors for automatically selecting web text most suitable for language modeling, a process typically referred to as quality filtering. Using a new dataset of U.S. high school newspaper articles -- written by students from across the country -- we investigate whose language is preferred by the quality filter used for GPT-3. We find that newspapers from larger schools, located in wealthier, educated, and urban ZIP codes are more likely to be classified as high quality. We then demonstrate that the filter's measurement of quality is unaligned with other sensible metrics, such as factuality or literary acclaim. We argue that privileging any corpus as high quality entails a language ideology, and more care is needed to construct training corpora for language models, with better transparency and justification for the inclusion or exclusion of various texts.