Elamur Research⌕
☰
home / papers / 2024 / KoBBQ: Korean Bias Benchmark for Question Answering

KoBBQ: Korean Bias Benchmark for Question Answering

Paper · 2024 · Jiho Jin, Jiseon Kim, Nayeon Lee et al.

Summary

Social bias benchmarks such as BBQ are built around US stereotypes, and simply machine-translating them misses how bias works in other societies. For example, drug use is linked to low socio-economic status in BBQ but to wealthy people in Korea. The authors propose a general pipeline for culturally adapting a BBQ-style dataset and apply it to Korean. Each BBQ sample is sorted into one of three classes:

  • Simply-Transferred: kept after culturally aware translation.
  • Target-Modified: the stereotype exists in Korea but applies to different groups, which are re-identified through a public survey.
  • Sample-Removed: dropped because it does not fit Korean society.

They also hand-write samples for four new Korea-specific categories. A large survey of Korean respondents checks every stereotype and its target group. The resulting benchmark, KoBBQ, is used to test six multilingual and Korean-specialized LLMs on QA accuracy and a diff-bias score. Model behavior on KoBBQ differs noticeably from behavior on machine-translated BBQ, which shows that hand-built, culturally grounded benchmarks are needed. The method can be reused for other languages and cultures.

Key points

  • KoBBQ contains 268 templates and 76,048 samples spanning 12 social bias categories. Each template yields context–question pairs in ambiguous or disambiguated contexts and in biased or counter-biased versions.
  • The Newly-Created samples cover four added categories: Educational Background, Political Orientation, Family Structure, and Domestic Area of Origin. Most tested LLMs showed high bias scores on these samples.
  • Validation asked 100 South Koreans per stereotype to pick a target group. A stereotype was discarded if over half saw no such stereotype or if responses were not skewed enough toward one group.
  • Unlike BBQ and CBBQ, KoBBQ is checked against a large-scale domestic survey. The dataset, evaluation code with prompts, and raw survey results are released at jinjh0123.github.io/KoBBQ.
  • The authors acknowledge a limitation: their bias metric is not fully separated from QA accuracy, so both should be read together.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Korea Advanced Institute of Science and TechnologyDaejeon · South Korea17 works · 2,673 citations🏢NAVER CloudSeongnam-si · South Korea13 works · 10,602 citations

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 9,086 worksArtificial Intelligence · 8,889 worksArts and Humanities · 9,296 worksLanguage and Linguistics · 9,206 worksSocial Sciences · 744 worksSafety Research · 221 works

🧭 Topics 3Research subjects and languages the work is about.Evaluation & benchmarks · 685 worksQuestion answering · 242 worksBias, fairness & safety · 209 works

Cite

Cite as Jiho Jin, Jiseon Kim, Nayeon Lee, Haneul Yoo, Alice Oh and Hwaran Lee. 2024. KoBBQ: Korean Bias Benchmark for Question Answering. Transactions of the Association for Computational Linguistics.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.1162/tacl_a_00661

@article{jin2024kobbq,
  title = {{KoBBQ: Korean Bias Benchmark for Question Answering}},
  author = {Jiho Jin, Jiseon Kim, Nayeon Lee, Haneul Yoo, Alice Oh and Hwaran Lee},
  journal = {Transactions of the Association for Computational Linguistics},
  year = {2024},
  doi = {10.1162/tacl\_a\_00661},
  url = {https://doi.org/10.1162/tacl_a_00661}
}

References

31 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Yuntao Bai, Saurav Kadavath, Sandipan Kundu, Amanda Askell, Jackson Kernion, Andy Jones, Anna Chen, Anna Goldie, Azalia Mirhoseini, Cameron McKinnon, Carol Chen, Catherine Olsson, Christopher Olah, Danny Hernandez, Dawn Drain, Deep Ganguli, Dustin Li, Eli Samuel Gehman, Suchin Gururangan, Maarten Sa…
  2. Association for Computational Linguistics. https://doi.org/10.18653/v1/2020 .findings-emnlp.301 — [doi:10.18653/v1/2020](https://doi.org/10.18653/v1/2020) ↗
  3. Kyung-Koo Han. 2007. The archaeology of the ethnically homogeneous nation-state and multiculturalism in Korea. Korea Journal, 47(4):8–32. https://doi.org/10.25024/kj.2007.47.4.8 ↗
  4. Yufei Huang and Deyi Xiong. 2023. CBBQ: A chinese bias benchmark dataset curated with human-ai collaboration for large language models. CoRR, abs/2306.16244v1. https://doi.org/10.48550/arXiv.2306.16244 ↗
  5. Hwaran Lee, Seokhee Hong, Joonsuk Park, Takyoung Kim, Gunhee Kim, and Jung-woo Ha. 2023a. KoSBI: A dataset for mitigating social bias risks towards safer large language model applications. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 5: Industry… find
  6. Nayeon Lee, Chani Jung, and Alice Oh. 2023b. Hate speech classifiers are culturally insensitive. In Proceedings of the First Workshop on Cross-Cultural Considerations in NLP (C3NLP), pages 35–46, Dubrovnik, Croatia. Association for Computational Linguistics. https://doi.org/10.18653/v1/2023 .c3nlp-1… ↗
  7. Kyung Min Im, Tae-Wan Kim, and Jong-Rok Jeon. 2017. Metal-chelation-assisted deposition of polydopamine on human hair: A ready-to-use eumelanin-based hair dyeing methodology. ACS Biomaterials Science & Engineering, 3(4):628–636. https://doi.org/10.1021/acsbiomaterials.7b00031, PubMed: 33429630 ↗
  8. Tao Li, Daniel Khashabi, Tushar Khot, Ashish Sabharwal, and Vivek Srikumar. 2020. UNQOVERing stereotyping biases via underspecified questions. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 3475–3489, Online. Association for Computational Linguistics. https://doi .or…
  9. Gautier Izacard, Patrick Lewis, Maria Lomeli, Lucas Hosseini, Fabio Petroni, Timo Schick, Jane Dwivedi-Yu, Armand Joulin, Sebastian Riedel, and Edouard Grave. 2023. Atlas: Few-shot learning with retrieval augmented language models. Journal of Machine Learning Research, 24(251):1–43. find
  10. Younghoon Jeong, Juhyun Oh, Jongwon Lee, Jaimeen Ahn, Jihyung Moon, Sungjoon Park, and Alice Oh. 2022. KOLD: Korean offensive language dataset. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 10818–10833, Abu Dhabi, United Arab Emirates. Association f… find
  11. Hadas Kotek, Rikker Dockum, and David Sun. 2023. Gender bias and stereotypes in large language models. In Proceedings of The ACM Collective Intelligence Conference, CI ’23, pages 12–24, New York, NY, USA. Association for Computing Machinery. https://doi.org/10.1145/3582269.3615599 — [doi:10.1145/3582269.3615599](https://doi.org/10.1145/3582269.3615599) ↗
  12. Ha-Ryoung Lee. 2007. Study on Social Prejudice towards Race: Centering on the Relationship of Social Distance to Stereotypes and emotions. Master’s thesis, Hanyang University, Seoul, KR. find
  13. Percy Liang, Rishi Bommasani, Tony Lee, Dimitris Tsipras, Dilara Soylu, Michihiro Yasunaga, Yian Zhang, Deepak Narayanan, Yuhuai Wu, Ananya Kumar, Benjamin Newman, Binhang Yuan, Bobby Yan, Ce Zhang, Christian Alexander Cosgrove, Christopher D. Manning, Christopher Re, Diana Acosta-Navas, Drew Arad H… find
  14. Bill Yuchen Lin, Seyeon Lee, Xiaoyang Qiao, and Xiang Ren. 2021. Common sense beyond English: Evaluating and improving multilingual language models for commonsense reasoning. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint C…
  15. Xuan-Phi Nguyen, Wenxuan Zhang, Xin Li, Mahani Aljunied, Qingyu Tan, Liying Cheng, Guanzheng Chen, Yue Deng, Sen Yang, Chaoqun Liu, Hang Zhang, and Lidong Bing. 2023. Seallms – large language models for southeast asia. https://doi.org/10.48550/arXiv.2312.00738 ↗
  16. Bill Yuchen Lin, Frank F. Xu, Kenny Zhu, and Seung-won Hwang. 2018. Mining cross-cultural differences and similarities in social media. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 709–719, Melbourne, Australia. Association… find
  17. Alicia Parrish, Angelica Chen, Nikita Nangia, Vishakh Padmakumar, Jason Phang, Jana Thompson, Phu Mon Htut, and Samuel Bowman. 2022. BBQ: A hand-built bias benchmark for question answering. In Findings of the Association for Computational Linguistics: ACL 2022, pages 2086–2105, Dublin, Ireland. Asso…
  18. Fangyu Liu, Emanuele Bugliarello, Edoardo Maria Ponti, Siva Reddy, Nigel Collier, and Desmond Elliott. 2021. Visually grounded reasoning across languages and cultures. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 10467–10485, Online and Punta Cana,… find
  19. Denis Peskov, Viktor Hangya, Jordan Boyd-Graber, and Alexander Fraser. 2021. Adapting entities across languages and cultures. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 3725–3750, Punta Cana, Dominican Republic. Association for Computational Linguistics. https://… find
  20. Fabio Motoki, Valdemar Pinho Neto, and Victor Rodrigues. 2023. More human than human: measuring chatgpt political bias. Public Choice. https://doi.org/10.1007/s11127-023-01097-2 ↗
  21. Edoardo Maria Ponti, Goran Glavaš, Olga Majewska, Qianchu Liu, Ivan Vulić, and Anna Korhonen. 2020. XCOPA: A multilingual dataset for causal commonsense reasoning. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 2362–2376, Online. Associatio…
  22. Moin Nadeem, Anna Bethke, and Siva Reddy. 2021. 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…
  23. Rachel Rudinger, Jason Naradowsky, Brian Leonard, and Benjamin Van Durme. 2018. 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), Nikita Na…
  24. Dilyar Buzan, Dimitri Coelho Mollo, Diyi Yang, Dong-Ho Lee, Dylan Schrader, Ekaterina Shutova, Ekin Dogus Cubuk, Elad Segal, Eleanor Hagerman, Elizabeth Barnes, Elizabeth Donoway, Ellie Pavlick, Emanuele Rodolà, Emma Lam, Eric Chu, Eric Tang, Erkut Erdem, Ernie Chang, Ethan A. Chi, Ethan Dyer, Etha… — [doi:10.1177/0022022194254006](https://doi.org/10.1177/0022022194254006) find
  25. Aarohi Srivastava, Abhinav Rastogi, Abhishek Rao, Abu Awal Md Shoeb, Abubakar Abid, Adam Fisch, Adam R. Brown, Adam Santoro, Aditya Gupta, Adrià Garriga-Alonso, Agnieszka Kluska, Aitor Lewkowycz, Akshat Agarwal, Alethea Power, Alex Ray, Alex Warstadt, Alexander W. Kocurek, Ali Safaya, Ali Tazarv, A…
  26. Yan Tao, Olga Viberg, Ryan S. Baker, and Rene F. Kizilcec. 2023. Auditing and mitigating cultural bias in llms. https://doi.org/10.48550/arXiv.2311.14096 ↗
  27. Jean-Paul Vinay and Jean Darbelnet. 1995. Comparative Stylistics of French and English: A methodology for translation. John Benjamins. https://doi.org/10.1075/btl.11 — [doi:10.1075/btl.11](https://doi.org/10.1075/btl.11) ↗
  28. Mingfeng Xue, Dayiheng Liu, Kexin Yang, Guanting Dong, Wenqiang Lei, Zheng Yuan, Chang Zhou, and Jingren Zhou. 2023. Occu-quest: Mitigating occupational bias for inclusive large language models. https://doi.org/10.48550/arXiv.2310.16517 — [doi:10.48550/arxiv.2310.16517](https://doi.org/10.48550/arxiv.2310.16517) ↗
  29. Jingfeng Yang, Haoming Jiang, Qingyu Yin, Danqing Zhang, Bing Yin, and Diyi Yang. 2022. SEQZERO: Few-shot compositional semantic parsing with sequential prompts and zero-shot models. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 49–60, Seattle, United States. Associ… find
  30. Da Yin, Liunian Harold Li, Ziniu Hu, Nanyun Peng, and Kai-Wei Chang. 2021. Broaden the vision: Geo-diverse visual commonsense reasoning. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 2115–2129, Online and Punta Cana, Dominican Republic. Association… find
  31. Jieyu Zhao, Tianlu Wang, Mark Yatskar, Vicente Ordonez, and Kai-Wei Chang. 2018. Gender bias in coreference resolution: Evaluation and debiasing methods. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies,…

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/jin-2024-kobbq-korean-bias-benchmark-question/