French CrowS-Pairs: Extending a challenge dataset for measuring social bias in masked language models to a language other than English
Open accessThe DOI matches its Crossref record; 9 of the 9 references it cites were found in a registry or an index.
Abstract
Warning: This paper contains explicit statements of offensive stereotypes which may be upsetting Much work on biases in natural language processing has addressed biases linked to the social and cultural experience of English speaking individuals in the United States.We seek to widen the scope of bias studies by creating material to measure social bias in language models (LMs) against specific demographic groups in France.We build on the US-centered CrowS-pairs dataset to create a multilingual stereotypes dataset that allows for comparability across languages while also characterizing biases that are specific to each country and language.We introduce 1,677 sentence pairs in French that cover stereotypes in ten types of bias like gender and age.1,467 sentence pairs are translated from CrowS-pairs and 210 are newly crowdsourced and translated back into English.The sentence pairs contrast stereotypes concerning underadvantaged groups with the same sentence concerning advantaged groups.We find that four widely used language models (three French, one multilingual) favor sentences that express stereotypes in most bias categories.We report on the translation process, which led to a characterization of stereotypes in CrowS-pairs including the identification of US-centric cultural traits.We offer guidelines to further extend the dataset to other languages and cultural environments.
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 · cited 10×
“We used the same protocol to evaluate the original three language models addressed by Nangia et al. (2020) as well as multilingual BERT.”
MethodInfluentialOn 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 2×
“The increasing use of large language models has raised many ethical concerns, including the risk of bias and bias amplification (Bender et al., 2021).”
BackgroundALBERT: A Lite BERT for Self-supervised Learning of Language Representations
Lan, Z., Chen, M., Goodman, S., Gimpel, K., Sharma, P., & Soricut, R. (2019) · arXiv · doi:10.48550/arxiv.1909.11942 · Cited by 4,308 · cited 7×
“5 Due to space constraints, we do not show results obtained for AlBERT large-v2 but they are consistent with the description provided in the previous paragraph.”
BackgroundInfluentialSemantics derived automatically from language corpora contain human-like biases
Caliskan, A., Bryson, J. J., & Narayanan, A. (2017) · Science · doi:10.1126/science.aal4230 · Cited by 2,887
Language (Technology) is Power: A Critical Survey of “Bias” in NLP
Blodgett, S. L., Barocas, S., Daumé III, H., & Wallach, H. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.485 · Cited by 1,050 · cited 1×
“Biases in NLP have received a lot of attention in recent years (Blodgett et al., 2020).”
BackgroundCrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Models
Nangia, N., Vania, C., Bhalerao, R., & Bowman, S. R. (2020) · Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.18653/v1/2020.emnlp-main.154 · Cited by 446
The social impact of natural language processing
Hovy, D., & Spruit, S. L. (2016) · Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) · doi:10.18653/v1/p16-2096 · Cited by 329 · cited 1×
“The natural language processing community who contributes to the development of these technologies has a responsibility to understand the social impact of its research and to address the ethical implications (Hovy and Spruit, 2016).”
BackgroundStereotyping Norwegian Salmon: An Inventory of Pitfalls in Fairness Benchmark Datasets
Blodgett, S. L., Lopez, G., Olteanu, A., Sim, R., & Wallach, H. (2021) · Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) · doi:10.18653/v1/2021.acl-long.81 · Cited by 147 · cited 2×
“During the translation process, we identified issues with the original CrowS-pairs content, including some described in Blodgett et al. (2021).”
BackgroundMethodUnmasking Contextual Stereotypes: Measuring and Mitigating BERT's Gender Bias
Bartl, M., Nissim, M., & Gatt, A. (2020) · arXiv · doi:10.48550/arxiv.2010.14534 · Cited by 84 · cited 1×
“A recent study focusing on gender bias in English and German has shown that methods to evidence and mitigate bias in English do not necessarily carry well to other languages (Bartl et al., 2020).”
Background
Cited by
Biases in Large Language Models: Origins, Inventory, and Discussion
Navigli, R., Conia, S., & Ross, B. (2023) · Journal of Data and Information Quality · doi:10.1145/3597307 · Cited by 364
“Only recently have comparisons of different fairness measures been carried out [40], and datasets of different types of social bias in English [95] and French [100] have also been made available.”
BackgroundHaving Beer after Prayer? Measuring Cultural Bias in Large Language Models
Naous, T., Ryan, M. J., Ritter, A., & Xu, W. (2024) · Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2024.acl-long.862 · Cited by 96
WinoQueer: A community-in-the-loop benchmark for anti-LGBTQ+ bias in large language models
Felkner, V., Chang, H.-C. H., Jang, E., & May, J. (2023) · Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · doi:10.18653/v1/2023.acl-long.507 · Cited by 74
Quantifying gender bias towards politicians in cross-lingual language models
Stańczak, K., Ray Choudhury, S., Pimentel, T., Cotterell, R., & Augenstein, I. (2023) · PLOS ONE · doi:10.1371/journal.pone.0277640 · Cited by 13
“We present a fine-grained study of gender bias in 6 cross-lingual language models.”
MethodLessons from the Trenches on Reproducible Evaluation of Language Models
Biderman, S., Schoelkopf, H., Sutawika, L., Gao, L., Tow, J., Abbasi, B., Aji, A. F., Ammanamanchi, P. S., Black, S., Clive, J., et al. (2024) · arXiv · doi:10.48550/arxiv.2405.14782 · Cited by 9
“…evaluations, especially more artificially-constructed benchmarks such as minimal-pair or pronoun-frequency (Nangia et al., 2020; Zhao et al., 2018; N´ev´eol et al., 2022) based bias benchmarks or multiple-choice knowledge tests (Hendrycks et al., 2020), provide information about how models will…”
Metadata
- Authors
- Névéol, A., Dupont, Y., Bezançon, J., & Fort, K.
- Venue
- Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
- Published
- 2022
- Pages
- 8521-8531
- OpenAlex
- W4281621415
- Cited by
- 38
- References
- 9
- FWCI
- 8.73
- 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.