@inproceedings{deas-etal-2025-data,
    title = "Data Caricatures: On the Representation of {A}frican {A}merican Language in Pretraining Corpora",
    author = "Deas, Nicholas  and
      Vente, Blake  and
      Ananthram, Amith  and
      Grieser, Jessica A  and
      Patton, Desmond U.  and
      Kleiner, Shana  and
      Iii, James R. Shepard  and
      McKeown, Kathleen",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingestion-acl-25/2025.acl-long.1416/",
    pages = "29192--29217",
    ISBN = "979-8-89176-251-0",
    abstract = "With a combination of quantitative experiments, human judgments, and qualitative analyses, we evaluate the quantity and quality of African American Language (AAL) representation in 12 predominantly English, open-source pretraining corpora. We specifically focus on the sources, variation, and naturalness of included AAL texts representing the AAL speaking community. We find that AAL is underrepresented in all evaluated pretraining corpora compared to US demographics, constituting as few as 0.007{\%} and at most 0.18{\%} of documents. We also find that more than 25{\%} of AAL texts in C4 may be perceived as inappropriate for LLMs to generate and to reinforce harmful stereotypes. Finally, we find that most automated filters are more likely to conserve White Mainstream English (WME) texts over AAL in pretraining corpora."
}