Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
41 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
41 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

In-Contextual Gender Bias Suppression for Large Language Models (2309.07251v2)

Published 13 Sep 2023 in cs.CL

Abstract: Despite their impressive performance in a wide range of NLP tasks, LLMs have been reported to encode worrying-levels of gender biases. Prior work has proposed debiasing methods that require human labelled examples, data augmentation and fine-tuning of LLMs, which are computationally costly. Moreover, one might not even have access to the model parameters for performing debiasing such as in the case of closed LLMs such as GPT-4. To address this challenge, we propose bias suppression that prevents biased generations of LLMs by simply providing textual preambles constructed from manually designed templates and real-world statistics, without accessing to model parameters. We show that, using CrowsPairs dataset, our textual preambles covering counterfactual statements can suppress gender biases in English LLMs such as LLaMA2. Moreover, we find that gender-neutral descriptions of gender-biased objects can also suppress their gender biases. Moreover, we show that bias suppression has acceptable adverse effect on downstream task performance with HellaSwag and COPA.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (54)
  1. Persistent anti-muslim bias in large language models. In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, pages 298–306.
  2. Evaluating gender bias of pre-trained language models in natural language inference by considering all labels. arXiv preprint arXiv:2309.09697.
  3. Stereotyping Norwegian salmon: An inventory of pitfalls in fairness benchmark datasets. In 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), pages 1004–1015, Online. Association for Computational Linguistics.
  4. Yang Trista Cao and Hal Daumé III. 2020. Toward gender-inclusive coreference resolution. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4568–4595, Online. Association for Computational Linguistics.
  5. Think you have solved question answering? try arc, the ai2 reasoning challenge. arXiv preprint arXiv:1803.05457.
  6. Harms of gender exclusivity and challenges in non-binary representation in language technologies. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 1968–1994, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
  7. Christiane Fellbaum. 2010. Wordnet. In Theory and applications of ontology: computer applications, pages 231–243. Springer.
  8. Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolution.
  9. The capacity for moral self-correction in large language models. arXiv preprint arXiv:2302.07459.
  10. He is very intelligent, she is very beautiful? on mitigating social biases in language modelling and generation. In Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 4534–4545.
  11. Xinyang Geng and Hao Liu. 2023. Openllama: An open reproduction of llama.
  12. Hila Gonen and Yoav Goldberg. 2019. Lipstick on a pig: Debiasing methods cover up systematic gender biases in word embeddings but do not remove them. In 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), pages 609–614, Minneapolis, Minnesota. Association for Computational Linguistics.
  13. Connecting large language models with evolutionary algorithms yields powerful prompt optimizers. arXiv preprint arXiv:2309.08532.
  14. Auto-debias: Debiasing masked language models with automated biased prompts. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1012–1023, Dublin, Ireland. Association for Computational Linguistics.
  15. Measuring massive multitask language understanding. arXiv preprint arXiv:2009.03300.
  16. Przemyslaw Joniak and Akiko Aizawa. 2022. Gender biases and where to find them: Exploring gender bias in pre-trained transformer-based language models using movement pruning. In Proceedings of the 4th Workshop on Gender Bias in Natural Language Processing (GeBNLP), pages 67–73, Seattle, Washington. Association for Computational Linguistics.
  17. Masahiro Kaneko and Danushka Bollegala. 2019. Gender-preserving debiasing for pre-trained word embeddings. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 1641–1650, Florence, Italy. Association for Computational Linguistics.
  18. Masahiro Kaneko and Danushka Bollegala. 2021a. Debiasing pre-trained contextualised embeddings. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 1256–1266.
  19. Masahiro Kaneko and Danushka Bollegala. 2021b. Dictionary-based debiasing of pre-trained word embeddings. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 212–223.
  20. Masahiro Kaneko and Danushka Bollegala. 2022. Unmasking the mask–evaluating social biases in masked language models. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pages 11954–11962.
  21. Debiasing isn’t enough! – on the effectiveness of debiasing MLMs and their social biases in downstream tasks. In Proceedings of the 29th International Conference on Computational Linguistics, pages 1299–1310, Gyeongju, Republic of Korea. International Committee on Computational Linguistics.
  22. Gender bias in masked language models for multiple languages. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2740–2750, Seattle, United States. Association for Computational Linguistics.
  23. Nora Kassner and Hinrich Schütze. 2020. Negated and misprimed probes for pretrained language models: Birds can talk, but cannot fly. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 7811–7818, Online. Association for Computational Linguistics.
  24. Sustainable modular debiasing of language models. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 4782–4797.
  25. Towards understanding and mitigating social biases in language models. In International Conference on Machine Learning, pages 6565–6576. PMLR.
  26. Truthfulqa: Measuring how models mimic human falsehoods. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 3214–3252.
  27. Few-shot learning with multilingual language models. arXiv preprint arXiv:2112.10668.
  28. It’s all in the name: Mitigating gender bias with name-based counterfactual data substitution. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 5267–5275, Hong Kong, China. Association for Computational Linguistics.
  29. Crosslingual generalization through multitask finetuning. arXiv preprint arXiv:2211.01786.
  30. Stereoset: Measuring stereotypical bias in pretrained language models. In 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), pages 5356–5371.
  31. Crows-pairs: A challenge dataset for measuring social biases in masked language models. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 1953–1967.
  32. Training language models to follow instructions with human feedback. Advances in Neural Information Processing Systems, 35:27730–27744.
  33. BBQ: A hand-built bias benchmark for question answering. In Findings of the Association for Computational Linguistics: ACL 2022, pages 2086–2105, Dublin, Ireland. Association for Computational Linguistics.
  34. Choice of plausible alternatives: An evaluation of commonsense causal reasoning. In AAAI spring symposium: logical formalizations of commonsense reasoning, pages 90–95.
  35. Gender bias in coreference resolution. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 2 (Short Papers), pages 8–14, New Orleans, Louisiana. Association for Computational Linguistics.
  36. Bloom: A 176b-parameter open-access multilingual language model. arXiv preprint arXiv:2211.05100.
  37. Self-diagnosis and self-debiasing: A proposal for reducing corpus-based bias in nlp. Transactions of the Association for Computational Linguistics, 9:1408–1424.
  38. The woman worked as a babysitter: On biases in language generation. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), pages 3407–3412, Hong Kong, China. Association for Computational Linguistics.
  39. Autoprompt: Eliciting knowledge from language models with automatically generated prompts. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 4222–4235.
  40. Prompting gpt-3 to be reliable. arXiv preprint arXiv:2210.09150.
  41. Gemini: a family of highly capable multimodal models. arXiv preprint arXiv:2312.11805.
  42. MosaicML NLP Team et al. 2023b. Introducing mpt-7b: A new standard for open-source, ly usable llms.
  43. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288.
  44. Language models don’t always say what they think: Unfaithful explanations in chain-of-thought prompting. arXiv preprint arXiv:2305.04388.
  45. Hrishikesh Viswanath and Tianyi Zhang. 2023. Fairpy: A toolkit for evaluation of social biases and their mitigation in large language models. arXiv preprint arXiv:2302.05508.
  46. Measuring and reducing gendered correlations in pre-trained models. arXiv preprint arXiv:2010.06032.
  47. Chain of thought prompting elicits reasoning in large language models. In Advances in Neural Information Processing Systems.
  48. Ronald J Williams and David Zipser. 1989. A learning algorithm for continually running fully recurrent neural networks. Neural computation, 1(2):270–280.
  49. Transformers: State-of-the-art natural language processing. In Proceedings of the 2020 conference on empirical methods in natural language processing: system demonstrations, pages 38–45.
  50. Hellaswag: Can a machine really finish your sentence? In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4791–4800.
  51. Gender bias in contextualized word embeddings. In 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), pages 629–634.
  52. Mengjie Zhao and Hinrich Schütze. 2021. Discrete and soft prompting for multilingual models. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 8547–8555, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics.
  53. Large language models are human-level prompt engineers. arXiv preprint arXiv:2211.01910.
  54. Counterfactual data augmentation for mitigating gender stereotypes in languages with rich morphology. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 1651–1661.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (3)
  1. Daisuke Oba (5 papers)
  2. Masahiro Kaneko (46 papers)
  3. Danushka Bollegala (84 papers)
Citations (3)
View on Semantic Scholar