The DOI matches its Crossref record; 34 of the 35 references it cites were found in a registry or an index.
CrowS-Pairs: A Challenge Dataset for Measuring Social Biases in Masked Language Models
Open accessAbstract
Warning: This paper contains explicit statements of offensive stereotypes and may be upsetting.Pretrained language models, especially masked language models (MLMs) have seen success across many NLP tasks.However, there is ample evidence that they use the cultural biases that are undoubtedly present in the corpora they are trained on, implicitly creating harm with biased representations.To measure some forms of social bias in language models against protected demographic groups in the US, we introduce the Crowdsourced Stereotype Pairs benchmark (CrowS-Pairs).CrowS-Pairs has 1508 examples that cover stereotypes dealing with nine types of bias, like race, religion, and age.In CrowS-Pairs a model is presented with two sentences: one that is more stereotyping and another that is less stereotyping.The data focuses on stereotypes about historically disadvantaged groups and contrasts them with advantaged groups.We find that all three of the widelyused MLMs we evaluate substantially favor sentences that express stereotypes in every category in CrowS-Pairs.As work on building less biased models advances, this dataset can be used as a benchmark to evaluate progress.
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
“In the literature, many works refer to the metric as the dataset, and use these interchangeably.”
From Bias and Fairness in Large Language Models: A Survey (2024) · cites this paper as Nangia et al. 2020
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
“Each pair consists of one sentence demonstrating a stereotype, and the other sentence replacing the disadvantaged group with an advantaged one (”
From Bias and Fairness in Large Language Models: A Survey (2024) · cites this paper as Nangia et al. 2020
Experimental — accuracy not yet measuredThe cited source says what the claim attributes to it.
In the citing paper
“InstructGPT does not significantly improve over GPT-3 on the Winogender (Rudinger et al., 2018) and CrowSPairs (Nangia et al., 2020) datasets.”
From Training language models to follow instructions with human feedback (2022) · cites this paper as Nangia et al., 2020
The cited source says what the claim attributes to it.
In the citing paper
“InstructGPT does not significantly improve over GPT-3 on the Winogender (Rudinger et al., 2018) and CrowSPairs (Nangia et al., 2020) datasets.”
From Training language models to follow instructions with human feedback (2022) · cites this paper as Nangia et al., 2020
The cited source says what the claim attributes to it.
In the citing paper
“One example is the CrowS-Pairs (Nangia et al. 2020) dataset consisting of stereotype sentence pairs proposed alongside a pseudo-log-likelihood metric over this specific set of sentence pairs:”
From Bias and Fairness in Large Language Models: A Survey (2024) · cites this paper as Nangia et al. 2020
The cited source says what the claim attributes to it.
In the citing paper
“The dataset covers nine types of biases: race, gender, sexual orientation, religion, age, nationality, disability, physical appearance, and socioeconomic status.”
From Bias and Fairness in Large Language Models: A Survey (2024) · cites this paper as Nangia et al. 2020
The cited source says what the claim attributes to it.
In the citing paper
“Crowdsourced Stereotype Pairs (CrowS-Pairs) (Nangia et al. 2020) provides 1,508 sentence pairs to evaluate stereotypes of historically disadvantaged social groups.”
From Bias and Fairness in Large Language Models: A Survey (2024) · cites this paper as Nangia et al. 2020
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
Glove: Global Vectors for Word Representation
Pennington, J., Socher, R., & Manning, C. (2014) · Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP) · doi:10.3115/v1/d14-1162 · Cited by 34,071
RoBERTa: A Robustly Optimized BERT Pretraining Approach
Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., & Stoyanov, V. (2019) · arXiv · doi:10.48550/arxiv.1907.11692 · Cited by 16,515
Deep Contextualized Word Representations
Peters, M., Neumann, M., Iyyer, M., Gardner, M., Clark, C., Lee, K., & Zettlemoyer, L. (2018) · Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers) · doi:10.18653/v1/n18-1202 · Cited by 11,983
ALBERT: 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
A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference
Williams, A., Nangia, N., & Bowman, S. (2018) · Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers) · doi:10.18653/v1/n18-1101 · Cited by 3,342
Self-citationSemantics 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
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
Know What You Don’t Know: Unanswerable Questions for SQuAD
Rajpurkar, P., Jia, R., & Liang, P. (2018) · Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) · doi:10.18653/v1/p18-2124 · Cited by 2,240
Black sexual politics: African Americans, gender, and the new racism
(2005) · Choice Reviews Online · doi:10.5860/choice.42-3729 · Cited by 1,940
Research works printed but not linked: 6
- Emily Dinan, Angela Fan, Ledell Wu, Jason Weston, Douwe Kiela, and Adina Williams. 2020. Multi- dimensional gender bias classification. ArXiv.
- Paul Pu Liang, Irene Mengze Li, Emily Zheng, Yao Chong Lim, Ruslan Salakhutdinov, and Louis- Philippe Morency. 2020. Towards debiasing sen- tence representations. In Proceedings of the 2020 Association for Computational Linguistics. Associa- tion for Computational Linguistics.
- Alex Wang and Kyunghyun Cho. 2019. BERT has a mouth, and it must speak: BERT as a Markov random field language model. In Proceedings of the Workshop on Methods for Optimizing and Eval- uating Neural Language Generation, pages 30-36, Minneapolis, Minnesota. Association for Computa- tional Linguistics. · DOI
- Alex Wang, Yada Pruksachatkun, Nikita Nangia, Amanpreet Singh, Julian Michael, Felix Hill, Omer Levy, and Samuel Bowman. 2019. SuperGLUE: A stickier benchmark for general-purpose language un- derstanding systems. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d Álché Buc, E. Fox, and R. Gar- nett, editors, Advances in Neural Information Pro- cessing Systems 32, pages 3266-3280.
- Kellie Webster, Marta Recasens, Vera Axelrod, and Ja- son Baldridge. 2018. Mind the GAP: A balanced corpus of gendered ambiguous pronouns. Transac- tions of the Association for Computational Linguis- tics, 6:605-617. · DOI · Paper page
- Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pier- ric Cistac, Tim Rault, Rémi Louf, Morgan Funtow- icz, and Jamie Brew. 2019. HuggingFace's trans- formers: State-of-the-art natural language process- ing. ArXiv.
Cited by
Sorted by citations
Items 401–456 of 456
Yang, K., Liu, D., & Wang, Z. (2026) · Synthesis lectures on human language technologies · doi:10.1007/978-3-032-06264-2_10
On Subjectivity, Bias and Fairness in Language Model Learning
Gaci, Y. (2023) · theses.fr (ABES) · doi:10.70675/dca82074za0c4z46d1z8a6bz96ce30bea3b6
Improving Representations for Language Modeling
Godey, N. (2024) · theses.fr (ABES) · doi:10.70675/b6a15d0ez15bez43e5za967z16a069e0a8bb
Sunitha, K., & Sharma, S. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-18346-0_23
SCOPE: A Dataset of Stereotyped Prompts for Counterfactual Fairness Assessment of LLMs
Parziale, A., Voria, G., Pontillo, V., Lucia, A. D., Catolino, G., & Palomba, F. (2026) · IEEE International Conference on Software Analysis, Evolution and Reengineering - Companion (SANER-C) · doi:10.1109/saner-c67878.2026.00025
Calibrating Stereotype Preference with Multi-Agent Graph Reasoning in Large Language Models
Amin, M. N., & Jesser, A. (2025) · International Conference on Computer and Information Technology (ICCIT) · doi:10.1109/iccit68739.2025.11491743
No LLM is Free From Bias: A Comprehensive Study of Bias Evaluation in Large Language Models
Charaka, V. K., Urlana, A., Kanumolu, G., Garlapati, B. M., & Mishra, P. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-19096-3_14 · Cited by 30
A Representation-Level Assessment of Bias Mitigation in Foundation Models
Nizhnichenkov, S., Nair, R., Daly, E. M., & Namee, B. M. (2026) · Communications in computer and information science · doi:10.1007/978-3-032-19096-3_13
Ge, Y., Gu, J., & Huang, C. (2026) · Data · doi:10.3390/data11060126
DetBias: A Framework for Detecting Bias in NLP Classifier Models
Adil, M. F., Aswani, S., & Shetty, S. D. (2026) · Lecture notes in networks and systems · doi:10.1007/978-3-032-25617-1_32
Don't be Tricked by Iterative Masking
Crothers, E., Viktor, H. L., & Japkowicz, N. (2026) · Computational Intelligence · doi:10.1111/coin.70247
Murdaugh, K. M., & Kwok, I. B. (2026) · Journal of Palliative Medicine · doi:10.1177/10966218261452780
The More You Say, the More You Risk: Ethical Concerns in Large Language Model Reasoning Frameworks
Zhao, J., Cao, L., Zhang, X., Shi, K., Wang, Y., & Penn, G. (2026) · International World Wide Web Conference (WWW) · doi:10.1145/3774905.3795475
Evaluating Social Bias in RAG Systems: When External Context Helps and Reasoning Hurts
Parihar, S., & Cheng, L. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-1468-6_44
Ahmed, A. M., Sadiq, A. T., Farhan, A. K., Ahmad, M., & Shakeel, M. R. (2026) · Archives of Computational Methods in Engineering · doi:10.1007/s11831-026-10656-x · Cited by 1
Expert Consensus on LLM Safety in Clinical Decision Support
Gayathri, K. V., Kanshiya, S., Saranraj, P., Lokeshwari, B., Vilmasen, P., & Wins, J. A. (2026) · Advances in computational intelligence and robotics book series · doi:10.4018/979-8-3373-7862-6.ch011
Open-Ended Bias Discovery in Large Language ModelsPreprint
Nguyen, T. H., & Luong, N. H. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6958779
Fu, W., Lu, Z., Qu, G., & Guo, X. (2026) · ACM Great Lakes Symposium on VLSI (GLSVLSI) · doi:10.1145/3787109.3815216
CoPAMoE: Cognitive Process-Aware Mixture of Experts for Bias Understanding and MitigationPreprint
Liu, X., Wang, F., Wang, F., Jia, Z., Wang, H., & Ran, C. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6987086
Gupta, A., Patil, N., Ghosh, S., & Gaikwad, S. S. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812394
CoPAMoE: Cognitive Process-Aware Mixture of Experts for Bias Understanding and MitigationPreprint
Liu, X., Wang, F., Wang, F., Jia, Z., Wang, H., & Ran, C. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.6997071
RobustDebias: Debiasing Language Models using Distributionally Robust Optimization
Gandhi, D., Singh, K., & Hegde, N. (2026) · ACM Conference on Fairness, Accountability, and Transparency (FAccT) · doi:10.1145/3805689.3812330
Nagaraj, P., Muneeswaran, V., Amer, A., Hafez, M., Muthiah, R., Islam, M. T., & Ijaz, M. F. (2025) · Asian Conference on Communication and Networks (ASIANComNet) · doi:10.1109/asiancomnet68615.2025.11579569
Xu, S., Zhang, L., Jamison, A., Saiyed, S., Hou, A. B., Du, H., & Gardner, L. (2026) · npj Digital Public Health · doi:10.1038/s44482-026-00026-6
Kurpicz-Briki, M., Ikae, C., Puttick, A. R., Rafnsdóttir, G. L., Haraldsdóttir, R. K., Rigotti, C., Fosch‐Villaronga, E., Sangiuliano, M., Kharas, M. W., & Søraa, R. A. (2026) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2026.1795293
Hong, C.-H., Liu, Y.-Z., & Chen, C. (2026) · Computers in Human Behavior Artificial Humans · doi:10.1016/j.chbah.2026.100350
Evaluating the Usage of African-American Vernacular English in Large Language Models
Davis, D., & McCoy, R. T. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-29903-1_15
Are Generative Recommenders Truly Fair? Auditing Bias with Conformal Calibration
Tao, C., Pan, Y., & Fan, Y. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-3444-8_22
Ferrari, M. (2026) · IGI Global eBooks · doi:10.4018/979-8-2600-3043-1.ch011
Understanding and Mitigating Competing Demographic Bias via Attention Disparity Analysis
Peng, T., Ci, R., Liu, Q., & Zhao, H. (2026) · Lecture notes in computer science · doi:10.1007/978-981-92-3417-2_7
Mitigating Bias in LLM-Augmented Knowledge Graphs: A Holistic Overview
Ibrahim, A., Kashef, R., & Naseribonari, S. (2026) · IEEE International Conference on Smart Sustainable Systems for Computer and Engineering Applications (3SCEA) · doi:10.1109/3scea68071.2026.11603048
BiasAgent: Exploiting Agent Bias for Preference Manipulation Attacks on Model Context Protocol
Guo, J. H., Wang, Z., Jiang, W., Zhang, R. Y., Xiong, J., Song, Q., Wu, H. B., & Liu, Y.-J. (2026) · IEEE Transactions on Cognitive Communications and Networking · doi:10.1109/tccn.2026.3714063
Wang, Y., & Steven, X. (2026) · International Conference on Artificial Intelligence and Digital Ethics (ICAIDE) · doi:10.1109/icaide69370.2026.11604541
SoK: Benchmark Datasets for Evaluating AI Safety — Gaps Between Guidelines and Practice
Ashizawa, N., & Saisho, O. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-32578-5_2
Tracing Stereotypes in Pre-trained Transformers: From Biased Neurons to Fairer Models
Voria, G., Openja, M., Khomh, F., Catolino, G., & Palomba, F. (2026) · International Conference on Mining Software Repositories (MSR) · doi:10.1145/3793302.3793356 · Cited by 3
FairBotBench: A Bilingual Benchmark for Evaluating Chatbot Behavior in E-commerce
Janusko, T., Bochnia, R., Hempel, G., Tomschke, S., Anke, J., & Thiele, M. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-26717-7_8
Deb, K., Mukherjee, I., & Sanyal, D. K. (2026) · Innovations in Systems and Software Engineering · doi:10.1007/s11334-026-00651-7
Cicatiello, C., & Fusco, P. (2026) · Encyclopedia · doi:10.3390/encyclopedia6090182
A review of bias detection and fairness auditing techniques in LLMs
Kaveti, N. K., & Pattanshetti, T. R. (2026) · Discover Artificial Intelligence · doi:10.1007/s44163-026-02121-2
Knowledge-Grounded Adaptive Reflection for Reliable and Fair Educational LLM Agents
Pan, K., Bi, Z., Zhao, K., Wang, Z., Liu, Z., & Lou, J. (2026) · International Conference on Large Models and Intelligent Agent Technologies (LMIAT) · doi:10.1109/lmiat70705.2026.11682086
Technofeminist Human-Centric AI in the Context of Conversational Large Language Models
Balontia, M. J., & Iskandar, I. F. R. (2026) · International Conference on Information Management and Technology (ICIMTech) · doi:10.1109/icimtech202671078.2026.11698952
StereoSet-PT: The Stereotype Evaluation Framework for Brazilian Portuguese Language Models
Aleixo, D. S., Costa, R. P., Pellicer, L. F. A. O., & de Almeida, R. A. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-39882-6_46
Pavlíček, M., Štěpánek, L., Kotyrba, M., & Volná, E. (2026) · Machine Learning and Knowledge Extraction · doi:10.3390/make8100303
Abhijeet, K., Ajay, M., Reddy, M. L. N., Santosh, D. T., & S, R. M. (2026) · International Conference on Emerging Techniques in Computational Intelligence (ICETCI) · doi:10.1109/icetci71772.2026.11710772
Beyond Static Guardrails: Auditable Surgical Output Control for Agentic LLM Workflows
Mauri, L., & Damiani, E. (2026) · IEEE International Conference on Web Services (ICWS) · doi:10.1109/icws72778.2026.11718616
Parziale, A., D'Auria, A., Scaparra, D. P., & Pontillo, V. (2026) · doi:10.1145/3843782.3844655
Liu, W., & Zhang, Y. (2026) · Internet Research · doi:10.1108/intr-10-2025-1717
Stereotype Bias Detection and Mitigation in Slovak and Multilingual Language Models
Sokolová, Z., Harahus, M., Sokol, M., & Kupcová, E. (2026) · International Conference Radioelektronika · doi:10.1109/radioelektronika71189.2026.11726461
Fair-MAD:For FGKB-Guided Adjudication of Stereotype-Prone Questions
Dong, Z., Yu, Z., Zuo, J., & Zhu, E. (2026) · International Conference on Computer Science, Machine Learning and Agents (CSMLA) · doi:10.1109/csmla70906.2026.11724473
Architecture-Aware Evaluation of Intrinsic Gender Bias in Masked and Autoregressive Language Models
Dhiman, M., & Chhabra, A. (2026) · International Conference on Trends in Engineering Systems and Technologies (ICTEST) · doi:10.1109/ictest69239.2026.11722874
Cardoso, L. R., & Ribeiro, L. S. F. (2026) · Simpósio Brasileiro de Tecnologia da Informação e da Linguagem Humana (STIL) · doi:10.5753/stil.2026.26812
PrefixGuard: Lightweight LLMs for Hate Speech and Sexism Detection
Dorleon, G., & Shujaa, S. (2026) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn69854.2026.11729560
AUTOMATED EVALUATION OF ETHICAL ASPECTS IN LARGE LANGUAGE MODELS
Kolchenko, V., & Sabodashko, D. (2026) · Cybersecurity Education Science Technique · doi:10.28925/2663-4023.2026.34.1357
Oraz, A. (2026) · Bangladesh Journal of Multidisciplinary Scientific Research · doi:10.46281/bjmsr.v12i1.2990
Serçe, A., Kaçıkan, E., & Onan, A. (2026) · International Artificial Intelligence and Data Processing Symposium (IDAP) · doi:10.1109/idap71869.2026.11727894
Behavioral Calibration for Context-Adaptive Debiasing in Large Language ModelsPreprint
Bai, Y., Gao, M., Wang, Z., Gu, B., Xiong, Q., Wang, X., & Yang, L. (2026) · SSRN Electronic Journal · doi:10.2139/ssrn.7675451
Metadata
- Authors
- Nangia, N., Vania, C., Bhalerao, R., & Bowman, S. R.
- Affiliations
- New York University (US)
- ORCID
- Bhalerao, Rasika
- Venue
- Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP)
- Published
- 2020
- Pages
- 1953-1967
- OpenAlex
- W3089430725
- Cited by
- 446
- References
- 35
- Citations by year
- 2026: 71 · 2025: 93 · 2024: 64 · 2023: 121 · 2022: 60 · 2021: 36 · 2020: 1
- FWCI
- 56.53
- Type
- conference paper
- Language
- en
- Open access
- hybrid
- doi.org
- Licence
- cc-by
- Locations
- Conference on Empirical Methods in Natural Language Processing (EMNLP) (publishedVersion, cc-by)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- TIB Data Manager
- Funders
- National Science Foundation (1922658)
- Nvidia
- Samsung
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
- Physical Sciences › Computer Science › Artificial Intelligence › Hate Speech and Cyberbullying Detection
- Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling
- Keywords
- social bias, masked language models, stereotypes, benchmark dataset, race, age, crowdsourced data
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
