Elamur Research⌕
☰
home / papers / 2026 / Do Political Opinions Transfer Between Western Languages? An Analysis of Unaligned and Aligned Multilingual LLMs

Do Political Opinions Transfer Between Western Languages? An Analysis of Unaligned and Aligned Multilingual LLMs

Paper · 2026 · Franziska Weeber, Tanise Ceron and Sebastian Padó

Summary

This paper asks whether multilingual LLMs hold a single, shared set of political views across languages or separate views per language, as human survey data across European countries might suggest. The authors tested 15 instruction-tuned bi- or multilingual models in German, English, French, Spanish and Italian. They used ProbVAA, a set of voting-advice-application statements from Ceron et al. (2024). Following that framework, they first discarded models whose stances changed under prompt rewording, then compared the agreement patterns of the remaining models across languages and policy areas. To tell genuine transfer apart from coincidental agreement, they used DPO to push two models from different families toward left or right positions, training only on English party-manifesto data. They then re-evaluated all five languages and added an open-ended statement-generation check. Before alignment, the models showed hardly any significant differences between languages. After alignment, opinions moved in nearly the same way in every language. The authors conclude that political opinions spread across Western languages inside these models. This makes alignment to a specific language or culture difficult and calls for cross-lingual evaluation.

Key points

  • The study covers 15 unaligned instruction-tuned MLLMs, mostly smaller models, in five Western European languages. Each model was screened for robustness before any opinion analysis.
  • ProbVAA statements are grouped into eight policy issues. Four are left-coded (environmental protection, welfare state, liberal society, open foreign policy). The other four are right-coded: tighter migration policy, fiscal restraint, economic liberalism, and an emphasis on law and order.
  • Political alignment used DPO with English manifestos only, applied to two models. The resulting left or right shift appeared almost evenly in all five languages, which supports the opinion-transfer hypothesis (H1) over separate per-language opinions (H2).
  • The authors note several limitations:
    • Only high-resource Western languages were tested.
    • Evaluation is mostly closed-ended, with only a small open-ended check.
    • Alignment was done in English only.
    • Effects on general capabilities after alignment were not tested.
  • Code is available at github.com/frawee/cross_lingual_opinion_transfer, and model responses and annotations are on OSF. The aligned models are deliberately not released because of misuse risk.

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

🎓University of StuttgartStuttgart · Germany91 works · 46,376 citations🎓Bocconi UniversityMilan · Italy28 works · 3,077 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksSocial Sciences · 1,772 worksSafety Research · 543 works

🧭 Topics 2Research subjects and languages the work is about.Language models & pretraining · 3,320 worksMultilingual & cross-lingual NLP · 828 works

Cite

Cite as Franziska Weeber, Tanise Ceron and Sebastian Padó. 2026. Do Political Opinions Transfer Between Western Languages? An Analysis of Unaligned and Aligned Multilingual LLMs. Proceedings of the 19th Conference of the European Chapter 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/v1/2026.eacl-long.14

@inproceedings{weeber2026do,
  title = {{Do Political Opinions Transfer Between Western Languages? An Analysis of Unaligned and Aligned Multilingual LLMs}},
  author = {Franziska Weeber, Tanise Ceron and Sebastian Padó},
  booktitle = {Proceedings of the 19th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.eacl-long.14},
  url = {https://doi.org/10.18653/v1/2026.eacl-long.14}
}

References

41 references, 16 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Marah Abdin, Jyoti Aneja, Hany Awadalla, Ahmed Awadallah, Ammar Ahmad Awan, Nguyen Bach, Amit Bahree, Arash Bakhtiari, Jianmin Bao, and Harkirat Behl et al. 2024. Phi-3 Technical Report: A Highly Capable Language Model Locally on Your Phone. arXiv:1234.56789.
  2. Ahmed Agiza, Mohamed Mostagir, and Sherief Reda. 2024. PoliTune: Analyzing the Impact of Data Selection and Fine-Tuning on Economic and Political Biases in Large Language Models. Proceedings of the AAAI/ACM Conference on AI, Ethics, and Society, 7(1):2–12. find
  3. Eleftherios Avramidis, Annika Grützner-Zahn, Manuel Brack, Patrick Schramowski, Pedro Ortiz Suarez, Malte Ostendorff, Fabio Barth, Shushen Manakhimova, Vivien Macketanz, Georg Rehm, and Kristian Kersting. 2024. Occiglot at WMT24: European Open-source Large Language Models Evaluated on Translation. I… find
  4. Tanise Ceron, Neele Falk, Ana Barić, Dmitry Nikolaev, and Sebastian Padó. 2024. Beyond Prompt Brittleness: Evaluating the Reliability and Consistency of Cohere. 2024. The Command R Model (Details and Application). Technical report. find
  5. Viviana Condorelli, Fiorenza Beluzzi, and Guido Anselmi. 2024. Assessing ChatGPT Political Bias in Italian Language. A Systematic Approach. Comunicazione politica, 3. find
  6. Julen Etxaniz, Gorka Azkune, Aitor Soroa, Oier Lopez de Lacalle, and Mikel Artetxe. 2024. Do Multilingual Language Models Think Better in English? In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume…
  7. 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…
  8. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, and Alex Vaughan et al. 2024. The Llama 3 Herd of Models. arXiv:2407.21783.
  9. Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2022. Lora: Low-rank adaptation of large language models. In The Tenth International Conference on Learning Representations, Online.
  10. Katharina Hämmerl, Jindřich Libovický, and Alexander Fraser. 2024. Understanding Cross-Lingual Alignment—A Survey. In Findings of the Association for Computational Linguistics: ACL 2024, pages 10922–10943, Bangkok, Thailand. Association for Computational Linguistics. find
  11. Albert Q. Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, and Lucie Saulnier et al. 2023. Mistral 7B. arXiv:2310.06825. find
  12. Albert Q. Jiang, Alexandre Sablayrolles, Antoine Roux, Arthur Mensch, Blanche Savary, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Emma Bou Hanna, and Florian Bressand et al. 2024. Mixtral of Experts. arXiv:2401.04088. find
  13. Klaus Krippendorff. 2019. Content Analysis: An Introduction to Its Methodology. SAGE Publications, Inc., Thousand Oaks, CA. find
  14. William H. Kruskal and W. Allen Wallis. 1952. Use of Ranks in One-Criterion Variance Analysis. Journal of the American Statistical Association, 47(260). find
  15. Pola Lehmann, Tobias Burst, Theres Matthieß, Sven Regel, Andrea Volkens, Bernhard Weßels, and Lisa Zehnter. 2022. The Manifesto Data Collection. Manifesto Project (MRG/CMP/MARPOR). find
  16. Version 2022a. Wissenschaftszentrum Berlin für Sozialforschung, Berlin. find
  17. Sharon Levy, Neha John, Ling Liu, Yogarshi Vyas, Jie Ma, Yoshinari Fujinuma, Miguel Ballesteros, Vittorio Castelli, and Dan Roth. 2023. Comparing Biases and the Impact of Multilingual Training across Multiple Languages. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language P… find
  18. Yifei Liu, Yuang Panwang, and Chao Gu. 2025. “Turning right”? An experimental study on the political value shift in large language models. Humanities and Social Sciences Communications, 12(1):179. find
  19. Nikita Nangia, Clara Vania, Rasika Bhalerao, and Samuel R. Bowman. 2020. 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, Online. Association for…
  20. Tarek Naous, Michael Ryan, Alan Ritter, and Wei Xu. 2024. Having Beer after Prayer? Measuring Cultural Bias in Large Language Models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 16366–16393, Bangkok, Thailand. Association…
  21. Vera Neplenbroek, Arianna Bisazza, and Raquel Fernández. 2024. MBBQ: A dataset for cross-lingual comparison of stereotypes in generative llms. In Conference on Language Modeling (COLM) 2024, Philadelphia, PA. find
  22. Vera Neplenbroek, Arianna Bisazza, and Raquel Fernández. 2025. Cross-lingual transfer of debiasing and detoxification in multilingual llms: An extensive investigation. arXiv:2412.14050. find
  23. Aurélie Névéol, Yoann Dupont, Julien Bezançon, and Karën Fort. 2022. French CrowS-Pairs: Extending a challenge dataset for measuring social bias in masked language models to a language other than English. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Vol…
  24. OpenAI. 2023. ChatGPT 3.5 turbo. Technical report. find
  25. 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…
  26. Yujin Potter, Shiyang Lai, Junsol Kim, James Evans, and Dawn Song. 2024. Hidden persuaders: LLMs’ political leaning and their influence on voters. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 4244– 4275, Miami, Florida, USA. Association for Computa… find
  27. Libo Qin, Qiguang Chen, Yuhang Zhou, Zhi Chen, Yinghui Li, Lizi Liao, Min Li, Wanxiang Che, and Philip S. Yu. 2025. A survey of multilingual large language models. Patterns, 6(1):101118. find
  28. Rafael Rafailov, Archit Sharma, Eric Mitchell, Stefano Ermon, Christopher D. Manning, and Chelsea Finn. 2024. Direct preference optimization: your language model is secretly a reward model. In Proceedings of the 37th International Conference on Neural Information Processing Systems, NIPS ’23, pages…
  29. Luca Rettenberger, Markus Reischl, and Mark Schutera. 2025. Assessing political bias in large language models. Journal of Computational Social Science, 8(2):42. find
  30. Paul Röttger, Musashi Hinck, Valentin Hofmann, Kobi Hackenburg, Valentina Pyatkin, Faeze Brahman, and Dirk Hovy. 2025. IssueBench: Millions of realistic prompts for measuring issue bias in LLM writing assistance. arXiv:2502.08395.
  31. David Rozado. 2024. The political preferences of LLMs. PLOS ONE, 19(7):e0306621. find
  32. Paul Röttger, Valentin Hofmann, Valentina Pyatkin, Musashi Hinck, Hannah Kirk, Hinrich Schuetze, and Dirk Hovy. 2024. Political Compass or Spinning Arrow? Towards More Meaningful Evaluations for Values and Opinions in Large Language Models. In Proceedings of the 62nd Annual Meeting of the Associatio…
  33. Shibani Santurkar, Esin Durmus, Faisal Ladhak, Cinoo Lee, Percy Liang, and Tatsunori Hashimoto. 2023. Whose Opinions Do Language Models Reflect? In Proceedings of the 40th International Conference on Machine Learning, pages 29971–30004. PMLR.
  34. Taylor Sorensen, Jared Moore, Jillian Fisher, Mitchell Gordon, Niloofar Mireshghallah, Christopher Michael Rytting, Andre Ye, Liwei Jiang, Ximing Lu, Nouha Dziri, Tim Althoff, and Yejin Choi. 2024. Position: a roadmap to pluralistic alignment. In Proceedings of the 41st International Conference on M…
  35. Dominik Stammbach, Philine Widmer, Eunjung Cho, Caglar Gulcehre, and Elliott Ash. 2024. Aligning Large Language Models with Diverse Political Viewpoints. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, pages 7257–7267, Miami, Florida, USA. Association for C… find
  36. Xiangpeng Wei, Haoran Wei, Huan Lin, Tianhao Li, Pei Zhang, Xingzhang Ren, Mei Li, Yu Wan, Zhiwei Cao, and Binbin Xie et al. 2023. PolyLM: An Open Source Polyglot Large Language Model. arXiv:.2307.06018. find
  37. Chris Wendler, Veniamin Veselovsky, Giovanni Monea, and Robert West. 2024. Do Llamas Work in English? On the Latent Language of Multilingual Transformers. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 15366–15394, Bangkok, T…
  38. Dustin Wright, Arnav Arora, Nadav Borenstein, Srishti Yadav, Serge Belongie, and Isabelle Augenstein. 2024. LLM Tropes: Revealing Fine-Grained Values and Opinions in Large Language Models. In Findings of the Association for Computational Linguistics: EMNLP 2024, pages 17085–17112, Miami, Florida, US… find
  39. Yuemei Xu, Ling Hu, Jiayi Zhao, Zihan Qiu, Kexin Xu, Yuqi Ye, and Hanwen Gu. 2025. A survey on multilingual large language models: corpora, alignment, and bias. Frontiers of Computer Science, 19(11):1911362. find
  40. Daniel M. Ziegler, Nisan Stiennon, Jeffrey Wu, Tom B. Brown, Alec Radford, Dario Amodei, Paul Christiano, and Geoffrey Irving. 2020. Fine-tuning language models from human preferences. arXiv:1909.08593. find
  41. Ahmet Üstün, Viraat Aryabumi, Zheng-Xin Yong, Wei-Yin Ko, Daniel D’souza, Gbemileke Onilude, Neel Bhandari, Shivalika Singh, Hui-Lee Ooi, and Amr Kayid et al. 2024. Aya Model: An Instruction Fine-tuned Open-Access Multilingual Language Model. arXiv:2402.07827.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/weeber-2026-do-political-opinions-transfer-between/