Elamur Research⌕
☰
home / papers / 2024 / Measuring Political Bias in Large Language Models: What Is Said and How It Is Said

Measuring Political Bias in Large Language Models: What Is Said and How It Is Said

Paper · 2024 · Yejin Bang, Delong Chen, Nayeon Lee et al.

Summary

Most bias benchmarks for LLMs focus on gender or race. When political leaning is studied, it is usually probed with human questionnaires such as the Political Compass, which reduce a model to one left–right score. Bang et al. instead examine what models actually write about contested issues. They prompt models to produce news headlines on topics such as reproductive rights, immigration, gun control, same-sex marriage, the death penalty and climate change.

A two-tier framework then analyses these outputs. First, it estimates stance per topic by comparing the model's generations with two extreme anchors, proponent and opponent texts that the same model writes on request. The imbalance in similarity to each anchor shows the direction and strength of the lean. Second, it examines framing along two axes: content (which frames and entities are selected) and style (biased writing style and sentiment toward key entities).

Applied to eleven open-source LLMs, the framework shows that political stance shifts from topic to topic. It offers an explainable, scalable alternative to survey-based audits.

Key points

  • Eleven open-source LLMs were evaluated on headline generation about politically sensitive topics. The authors note that the findings are tied to this task, though the method itself does not depend on the task.
  • Stance is measured with an anchor-based distance comparison. Framing is split into content, analysed with Boydstun's frame dimensions and entity-based frames, and style, analysed through media-bias writing style and sentiment toward salient entities.
  • Models tend to lean liberal on reproductive rights but more conservative on immigration. Overall they lean liberal, in line with earlier survey-based work.
  • Larger models are not necessarily more neutral, and models from the same family can show different biases. Multilingual models such as Yi-chat and Jais-chat differ in thematic focus from mainly English-trained models, and generations frequently centre on US contexts.
  • Stated limitations include the binary pro/opp anchors, a possible violation of the equal-norm assumption for self-generated anchors, and no treatment of hallucination. The code is released at github.com/HLTCHKUST/llm-political-bias.

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

🎓Hong Kong University of Science and TechnologyHong Kong · China142 works · 25,136 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksSocial Sciences · 1,772 worksSafety Research · 543 works

🧭 Topics 2Research subjects and languages the work is about.Language models & pretraining · 3,320 worksBias, fairness & safety · 517 works

Cite

Cite as Yejin Bang, Delong Chen, Nayeon Lee and Pascale Fung. 2024. Measuring Political Bias in Large Language Models: What Is Said and How It Is Said. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2024.acl-long.600

@inproceedings{bang2024measuring,
  title = {{Measuring Political Bias in Large Language Models: What Is Said and How It Is Said}},
  author = {Yejin Bang, Delong Chen, Nayeon Lee and Pascale Fung},
  booktitle = {Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2024},
  doi = {10.18653/v1/2024.acl-long.600},
  url = {https://doi.org/10.18653/v1/2024.acl-long.600}
}

References

75 references, 26 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Mohammad Ali and Naeemul Hassan. 2022. A survey of computational framing analysis approaches. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 9335–9348, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics. find
  2. Ebtesam Almazrouei, Hamza Alobeidli, Abdulaziz Alshamsi, Alessandro Cappelli, Ruxandra Cojocaru, Merouane Debbah, Etienne Goffinet, Daniel Heslow, Julien Launay, Quentin Malartic, Badreddine Noune, Baptiste Pannier, and Guilherme Penedo. 2023. Falcon-40B: an open large language model with state-of-t… find
  3. Lisa P. Argyle, Ethan C. Busby, Nancy Fulda, Joshua R. Gubler, Christopher Rytting, and David Wingate. 2023. Out of one, many: Using language models to simulate human samples. Political Analysis, 31(3):337–351.
  4. Ramy Baly, Giovanni Da San Martino, James Glass, and Preslav Nakov. 2020. We can detect your bias: Predicting the political ideology of news articles. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 4982–4991, Online. Association for Computati…
  5. Yejin Bang, Nayeon Lee, and Pascale Fung. 2023. Mitigating framing bias with polarity minimization loss. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 11100–11110, Singapore. Association for Computational Linguistics. find
  6. Yejin Bang, Nayeon Lee, Etsuko Ishii, Andrea Madotto, and Pascale Fung. 2021. Assessing political prudence of open-domain chatbots. In Proceedings of the 22nd Annual Meeting of the Special Interest Group on Discourse and Dialogue, pages 548– 555, Singapore and Online. Association for Computational L… find
  7. Adriana Beratšová, Kristína Krchová, Nikola Gažová, and Michal Jirásek. 2016. Framing and bias: a literature review of recent findings. Central European journal of management, 3(2). find
  8. Su Lin Blodgett, Solon Barocas, Hal Daumé III, and Hanna Wallach. 2020. Language (technology) is power: A critical survey of “bias” in NLP. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5454– 5476, Online. Association for Computational Linguistics.
  9. Amber E Boydstun, Dallas Card, Justin H Gross, Philip Resnik, and Noah A Smith. 2014. Tracking the development of media frames within and across policy issues. find
  10. Dallas Card, Amber E. Boydstun, Justin H. Gross, Philip Resnik, and Noah A. Smith. 2015. The media frames corpus: Annotations of frames across issues. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natural Lang…
  11. Tanise Ceron, Neele Falk, Ana Barić, Dmitry Nikolaev, and Sebastian Padó. 2024. Beyond prompt brittleness: Evaluating the reliability and consistency of political worldviews in llms.
  12. Wei-Fan Chen, Khalid Al Khatib, Henning Wachsmuth, and Benno Stein. 2020. Analyzing political bias and unfairness in news articles at different levels of granularity. In Proceedings of the Fourth Workshop on Natural Language Processing and Computational Social Science, pages 149–154, Online. Associa… find
  13. Wei-Fan Chen, Henning Wachsmuth, Khalid Al-Khatib, and Benno Stein. 2018. Learning to flip the bias of news headlines. In Proceedings of the 11th International Conference on Natural Language Generation, pages 79–88, Tilburg University, The Netherlands. Association for Computational Linguistics. find
  14. Dennis Chong and James N Druckman. 2007. Framing theory. Annu. Rev. Polit. Sci., 10:103–126. find
  15. Stefano DellaVigna and Ethan Kaplan. 2007. The fox news effect: Media bias and voting. The Quarterly Journal of Economics, 122(3):1187–1234. find
  16. Dorottya Demszky, Nikhil Garg, Rob Voigt, James Zou, Jesse Shapiro, Matthew Gentzkow, and Dan Jurafsky. 2019. Analyzing polarization in social media: Method and application to tweets on 21 mass shootings. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Comp… find
  17. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2018. BERT: pre-training of deep bidirectional transformers for language understanding. CoRR, abs/1810.04805.
  18. Jwala Dhamala, Tony Sun, Varun Kumar, Satyapriya Krishna, Yada Pruksachatkun, Kai-Wei Chang, and Rahul Gupta. 2021. Bold: Dataset and metrics for measuring biases in open-ended language generation. In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, FAccT ’21, pa… find
  19. Esin Durmus, Karina Nyugen, Thomas I Liao, Nicholas Schiefer, Amanda Askell, Anton Bakhtin, Carol Chen, Zac Hatfield-Dodds, Danny Hernandez, Nicholas Joseph, et al. 2023. Towards measuring the representation of subjective global opinions in language models. arXiv preprint arXiv:2306.16388.
  20. Erick Elejalde, Leo Ferres, and Eelco Herder. 2018. On the nature of real and perceived bias in the mainstream media. PloS one, 13(3):e0193765. find
  21. Robert M Entman. 1993. Framing: Toward clarification of a fractured paradigm. Journal of communication, 43(4):51–58. find
  22. Robert M Entman. 2007. Framing bias: Media in the distribution of power. Journal of communication, 57(1):163–173. find
  23. Lisa Fan, Marshall White, Eva Sharma, Ruisi Su, Prafulla Kumar Choubey, Ruihong Huang, and Lu Wang. 2019. In plain sight: Media bias through the lens of factual reporting. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Confer… find
  24. Shangbin Feng, Chan Young Park, Yuhan Liu, and Yulia Tsvetkov. 2023. From pretraining data to language models to downstream tasks: Tracking the trails of political biases leading to unfair NLP models. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume…
  25. Deep Ganguli, Liane Lovitt, Jackson Kernion, Amanda Askell, Yuntao Bai, Saurav Kadavath, Ben Mann, Ethan Perez, Nicholas Schiefer, Kamal Ndousse, et al. 2022. Red teaming language models to reduce harms: Methods, scaling behaviors, and lessons learned. arXiv preprint arXiv:2209.07858.
  26. R Kelly Garrett. 2009. Politically motivated reinforcement seeking: Reframing the selective exposure debate. Journal of communication, 59(4):676–699. find
  27. Matthew Gentzkow and Jesse M Shapiro. 2006. Media bias and reputation. Journal of political Economy, 114(2):280–316. find
  28. David Giles and Rachel L Shaw. 2009. The psychology of news influence and the development of media framing analysis. Social and personality psychology compass, 3(4):375–393. find
  29. Erving Goffman. 1974. Frame analysis: An essay on the organization of experience. Harvard University Press. find
  30. Joshua Gordon, Marzieh Babaeianjelodar, and Jeanna Matthews. 2020. Studying political bias via word embeddings. In Companion Proceedings of the Web Conference 2020, WWW ’20, page 760–764, New York, NY, USA. Association for Computing Machinery. find
  31. Felix Hamborg. 2020. Media bias, the social sciences, and NLP: Automating frame analyses to identify bias by word choice and labeling. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: Student Research Workshop, pages 79–87, Online. Association for Computati… find
  32. Felix Hamborg and Karsten Donnay. 2021. NewsMTSC: A dataset for (multi-)target-dependent sentiment classification in political news articles. In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, pages 1663–1675, Online. Associat… find
  33. Saghar Hosseini, Hamid Palangi, and Ahmed Hassan Awadallah. 2023. An empirical study of metrics to measure representational harms in pre-trained language models. In Proceedings of the 3rd Workshop on Trustworthy Natural Language Processing (TrustNLP 2023), pages 121–134, Toronto, Canada. Association…
  34. Ziwei Ji, Nayeon Lee, Rita Frieske, Tiezheng Yu, Dan Su, Yan Xu, Etsuko Ishii, Ye Jin Bang, Andrea Madotto, and Pascale Fung. 2023. Survey of hallucination in natural language generation. ACM Computing Surveys, 55(12):1–38.
  35. Hang Jiang, Doug Beeferman, Brandon Roy, and Deb Roy. 2022. CommunityLM: Probing partisan worldviews from language models. In Proceedings of the 29th International Conference on Computational Linguistics, pages 6818–6826, Gyeongju, Republic of Korea. International Committee on Computational Linguist… find
  36. Kristen Johnson and Dan Goldwasser. 2016. Identifying stance by analyzing political discourse on Twitter. In Proceedings of the First Workshop on NLP and Computational Social Science, pages 66–75, Austin, Texas. Association for Computational Linguistics. find
  37. Lalitha Kameswari and Radhika Mamidi. 2021. Towards quantifying magnitude of political bias in news articles using a novel annotation schema. In Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021), pages 671–678, Held Online. INCOMA Ltd. find
  38. Been Kim, Martin Wattenberg, Justin Gilmer, Carrie J. Cai, James Wexler, Fernanda B. Viégas, and Rory Sayres. 2018. Interpretability beyond feature attribution: Quantitative testing with concept activation vectors (TCAV). In Proceedings of the 35th International Conference on Machine Learning, ICML… find
  39. Dahyun Kim, Chanjun Park, Sanghoon Kim, Wonsung Lee, Wonho Song, Yunsu Kim, Hyeonwoo Kim, Yungi Kim, Hyeonju Lee, Jihoo Kim, Changbae Ahn, Seonghoon Yang, Sukyung Lee, Hyunbyung Park, Gyoungjin Gim, Mikyoung Cha, Hwalsuk Lee, and Sunghun Kim. 2023. Solar 10.7b: Scaling large language models with sim… find
  40. Sachin Kumar, Vidhisha Balachandran, Lucille Njoo, Antonios Anastasopoulos, and Yulia Tsvetkov. 2023. Language generation models can cause harm: So what can we do about it? an actionable survey. In Proceedings of the 17th Conference of the European Chapter of the Association for Computational Lingui…
  41. Nayeon Lee, Yejin Bang, Tiezheng Yu, Andrea Madotto, and Pascale Fung. 2022. NeuS: Neutral multi-news summarization for mitigating framing bias. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 31… find
  42. Ruibo Liu, Chenyan Jia, Jason Wei, Guangxuan Xu, and Soroush Vosoughi. 2022. Quantifying and alleviating political bias in language models. Artificial Intelligence, 304:103654. find
  43. Ruibo Liu, Chenyan Jia, Jason Wei, Guangxuan Xu, Lili Wang, and Soroush Vosoughi. 2021. Mitigating political bias in language models through reinforced calibration. In Proceedings of the AAAI Conference on Artificial Intelligence, pages 14857–14866. find
  44. Yang Liu, Yuanshun Yao, Jean-Francois Ton, Xiaoying Zhang, Ruocheng Guo Hao Cheng, Yegor Klochkov, Muhammad Faaiz Taufiq, and Hang Li. 2023. Trustworthy llms: a survey and guideline for evaluating large language models’ alignment. arXiv preprint arXiv:2308.05374.
  45. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. find
  46. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692.
  47. Fabio Motoki, Valdemar Pinho Neto, and Victor Rodrigues. 2024. More human than human: Measuring chatgpt political bias. Public Choice, 198(1):3–23. find
  48. 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…
  49. Ethan Perez, Sam Ringer, Kamile Lukosiute, Karina Nguyen, Edwin Chen, Scott Heiner, Craig Pettit, Catherine Olsson, Sandipan Kundu, Saurav Kadavath, Andy Jones, Anna Chen, Benjamin Mann, Brian Israel, Bryan Seethor, Cameron McKinnon, Christopher Olah, Da Yan, Daniela Amodei, Dario Amodei, Dawn Drain… — Discovering Language Model Behaviors with Model-Written Evaluations
  50. Daniel Preoţiuc-Pietro, Ye Liu, Daniel Hopkins, and Lyle Ungar. 2017. Beyond binary labels: Political ideology prediction of Twitter users. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 729–740, Vancouver, Canada. Associati… find
  51. Nils Reimers and Iryna Gurevych. 2019. Sentence-BERT: Sentence embeddings using Siamese BERT-networks. 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 3982–3992,…
  52. Paul Röttger, Valentin Hofmann, Valentina Pyatkin, Musashi Hinck, Hannah Rose Kirk, Hinrich Schütze, and Dirk Hovy. 2024. Political compass or spinning arrow? towards more meaningful evaluations for values and opinions in large language models. arXiv preprint arXiv:2402.16786.
  53. Shamik Roy and Dan Goldwasser. 2020. Weakly supervised learning of nuanced frames for analyzing polarization in news media. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 7698–7716, Online. Association for Computational Linguistics. find
  54. David Rozado. 2023. The political biases of chatgpt. Social Sciences, 12(3):148. find
  55. David Rozado. 2024. The political preferences of llms. find
  56. Diego Saez-Trumper, Carlos Castillo, and Mounia Lalmas. 2013. Social media news communities: gatekeeping, coverage, and statement bias. In Proceedings of the 22nd ACM international conference on Information & Knowledge Management, pages 1679– 1684. find
  57. Dietram A Scheufele. 1999. Framing as a theory of media effects. Journal of communication, 49(1):103– 122. find
  58. Patrick Schramowski, Cigdem Turan, Nico Andersen, Constantin A Rothkopf, and Kristian Kersting. 2022. Large pre-trained language models contain human-like biases of what is right and wrong to do. Nature Machine Intelligence, 4(3):258–268. find
  59. Neha Sengupta, Sunil Kumar Sahu, Bokang Jia, Satheesh Katipomu, Haonan Li, Fajri Koto, Osama Mohammed Afzal, Samta Kamboj, Onkar Pandit, Rahul Pal, et al. 2023. Jais and jais-chat: Arabic-centric foundation and instruction-tuned open generative large language models. arXiv preprint arXiv:2308.16149.
  60. Nikhil Sharma, Q Vera Liao, and Ziang Xiao. 2024. Generative echo chamber? effect of llm-powered search systems on diverse information seeking. In Proceedings of the CHI Conference on Human Factors in Computing Systems, pages 1–17. find
  61. Emily Sheng, Kai-Wei Chang, Prem Natarajan, and Nanyun Peng. 2021. Societal biases in language generation: Progress and challenges. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (V…
  62. Eric Michael Smith, Melissa Hall, Melanie Kambadur, Eleonora Presani, and Adina Williams. 2022. “I’m sorry to hear that”: Finding new biases in language models with a holistic descriptor dataset. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 9180–92…
  63. Irene Solaiman and Christy Dennison. 2021. Process for adapting language models to society (palms) with values-targeted datasets. Advances in Neural Information Processing Systems, 34:5861–5873. find
  64. Timo Spinde, Manuel Plank, Jan-David Krieger, Terry Ruas, Bela Gipp, and Akiko Aizawa. 2021. Neural media bias detection using distant supervision with BABE - bias annotations by experts. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 1166–1177, Punta Cana, Dominican… find
  65. Maximilian Spliethöver, Maximilian Keiff, and Henning Wachsmuth. 2022. No word embedding model is perfect: Evaluating the representation accuracy for social bias in the media. In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 2081–2093, Abu Dhabi, United Arab Emirates.… find
  66. Natalie Jomini Stroud. 2010. Polarization and partisan selective exposure. Journal of communication, 60(3):556–576. find
  67. Tianxiang Sun, Junliang He, Xipeng Qiu, and Xuanjing Huang. 2022. BERTScore is unfair: On social bias in language model-based metrics for text generation. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 3726–3739, Abu Dhabi, United Arab Emirates. Asso… find
  68. Erik F. Tjong Kim Sang and Fien De Meulder. 2003. Introduction to the CoNLL-2003 shared task: Language-independent named entity recognition. In Proceedings of the Seventh Conference on Natural Language Learning at HLT-NAACL 2003, pages 142– 147.
  69. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. 2023. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288.
  70. Dror Walter and Yotam Ophir. 2019. News frame analysis: An inductive mixed-method computational approach. Communication Methods and Measures, 13(4):248–266. find
  71. Yuxuan Wan, Wenxuan Wang, Pinjia He, Jiazhen Gu, Haonan Bai, and Michael R. Lyu. 2023. Biasasker: Measuring the bias in conversational ai system. In Proceedings of the 31st ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering, ESEC/FSE 2023, pag… find
  72. Laura Weidinger, John Mellor, Maribeth Rauh, Conor Griffin, Jonathan Uesato, Po-Sen Huang, Myra Cheng, Mia Glaese, Borja Balle, Atoosa Kasirzadeh, et al. 2021. Ethical and social risks of harm from language models. arXiv preprint arXiv:2112.04359.
  73. Yue Yu, Yuchen Zhuang, Jieyu Zhang, Yu Meng, Alexander Ratner, Ranjay Krishna, Jiaming Shen, and Chao Zhang. 2023. Large language model as attributed training data generator: A tale of diversity and bias. In Thirty-seventh Conference on Neural Information Processing Systems Datasets and Benchmarks T…
  74. Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric Xing, Hao Zhang, Joseph E. Gonzalez, and Ion Stoica. 2023. Judging LLM-as-a-judge with MT-bench and chatbot arena. In Thirty-seventh Conference on Neural Information Processing…
  75. Caleb Ziems and Diyi Yang. 2021. To protect and to serve? analyzing entity-centric framing of police violence. In Findings of the Association for Computational Linguistics: EMNLP 2021, pages 957–976, Punta Cana, Dominican Republic. Association for Computational Linguistics. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/bang-2024-measuring-political-bias-large-language/