Elamur Research⌕
☰
home / papers / 2023 / Knowledge of cultural moral norms in large language models

Knowledge of cultural moral norms in large language models

Paper · 2023 · Aida Ramezani and Yang Xu

Summary

Moral judgments differ between societies. Earlier studies found that English language models reflect human moral intuitions, but they assumed a single homogeneous set of norms. This paper tests whether monolingual English pre-trained language models (EPLMs) hold country-specific moral knowledge. It examines two questions: whether models track how acceptability of specific topics (e.g., divorce) differs across countries, and whether they capture which topics nations agree or disagree on.

The ground truth comes from World Values Survey wave 7 (55 countries) and PEW Global Attitudes (40 countries). The authors adapt the Schramowski et al. MORAL DIRECTION approach for SBERT by inserting country names into prompts. For GPT2 and GPT3, they introduce a new probe based on contrasting positive and negative moral completions.

Pre-trained models predicted cross-country norms less accurately than the English norms reported earlier, and they did better for western than non-western countries. Fine-tuning GPT2 on the survey data improved cross-cultural predictions but weakened its estimates of homogeneous English norms. The work highlights cultural bias risks for models deployed in multicultural settings.

Key points

  • Data: WVS wave 7 (55 countries) and PEW global morality survey (40 countries). Country is used as a proxy for culture.
  • Models: SBERT (bert-large-nli-mean-tokens), GPT2 (117M), GPT2-MEDIUM (345M), GPT2-LARGE (774M) and GPT3 (175B).
  • GPT probe: Prompts take the form "In [Country] [Topic]" followed by K = 5 opposing judgment pairs (e.g., right/wrong). The moral score is the log-probability difference of the final tokens, averaged over the pairs.
  • Fine-tuning results: GPT2 fine-tuned with a Random split reached Pearson's r of 0.893 (WVS) and 0.944 (PEW) for predicting cross-country diversity per topic. A western/non-western gap persisted, and performance on homogeneous norms dropped.
  • Resources: Code and data are released at github.com/AidaRamezani/cultural_inference.

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 TorontoToronto · Canada47 works · 50,254 citations

📚 Shelves 8The 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 worksNeuroscience · 95 worksCognitive Neuroscience · 95 works

🧭 Topics 1Research subjects and languages the work is about.Language models & pretraining · 1,343 works

Cite

Cite as Aida Ramezani and Yang Xu. 2023. Knowledge of cultural moral norms in large language models. Proceedings of the 61st 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/v1/2023.acl-long.26

@inproceedings{ramezani2023knowledge,
  title = {{Knowledge of cultural moral norms in large language models}},
  author = {Aida Ramezani and Yang Xu},
  booktitle = {Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2023},
  doi = {10.18653/v1/2023.acl-long.26},
  url = {https://doi.org/10.18653/v1/2023.acl-long.26}
}

References

41 references, 12 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Abubakar Abid, Maheen Farooqi, and James Zou. 2021. Persistent anti-muslim bias in large language models. In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society, AIES ’21, page 298–306, New York, NY, USA. Association for Computing Machinery.
  2. Prithviraj Ammanabrolu, Liwei Jiang, Maarten Sap, Hannaneh Hajishirzi, and Yejin Choi. 2022. Aligning to social norms and values in interactive narratives. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologie… find
  3. Arnav Arora, Lucie-Aimée Kaffee, and Isabelle Augenstein. 2022. Probing Pre-Trained Language Models for Cross-Cultural Differences in Values. arXiv preprint arXiv:2203.13722.
  4. Mohammad Atari, Jonathan Haidt, Jesse Graham, Sena Koleva, Sean Stevens, and Morteza Dehghani. 2022. Morality Beyond the WEIRD: How the Nomological Network of Morality Varies Across Cultures. the 2021 Conference on Empirical Methods in Natural Language Processing, pages 698–718, Online and Punta Can… find
  5. Maxwell Forbes, Jena D. Hwang, Vered Shwartz, Maarten Sap, and Yejin Choi. 2020. Social Chemistry 101: Learning to Reason about Social and Moral Norms. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 653–670, Online. Association for Computatio… find
  6. Jesse Graham, Jonathan Haidt, Sena Koleva, Matt Motyl, Ravi Iyer, Sean P Wojcik, and Peter H Ditto. 2013. Moral Foundations Theory: The Pragmatic Validity of Moral Pluralism. In Advances in Experimental Social Psychology, volume 47, pages 55–130. Elsevier. find
  7. Christian Haerpfer, Ronald Inglehart, Alejandro Moreno, Christian Welzel, Kseniya Kizilova, Jaime Diez-Medrano, Marta Lagos, Pippa Norris, E Ponarin, and B Puranen. 2021. World Values Survey: Round Seven – Country-Pooled Datafile. Madrid, Spain & Vienna, Austria: JD Systems Institute & WVSA Secretar… find
  8. Edmond Awad, Sohan Dsouza, Richard Kim, Jonathan Schulz, Joseph Henrich, Azim Shariff, Jean-François Bonnefon, and Iyad Rahwan. 2018. The Moral Machine experiment. Nature, 563(7729):59– 64. find
  9. Jonathan Haidt, Silvia Helena Koller, and Maria G Dias. 1993. Affect, culture, and morality, or is it wrong to eat your dog? Journal of personality and social psychology, 65(4):613. find
  10. Edmond Awad, Sohan Dsouza, Azim Shariff, Iyad Rahwan, and Jean-François Bonnefon. 2020. Universals and variations in moral decisions made in 42 countries by 70,000 participants. Proceedings of the National Academy of Sciences, 117(5):2332–2337. find
  11. Katharina Hämmerl, Björn Deiseroth, Patrick Schramowski, Jindřich Libovickỳ, Alexander Fraser, and Kristian Kersting. 2022. Do Multilingual Language Models Capture Differing Moral Norms? arXiv preprint arXiv:2203.09904. — [arXiv:2203.09904](https://arxiv.org/abs/2203.09904) find
  12. Cristina Bicchieri. 2005. The grammar of society: The nature and dynamics of social norms. Cambridge University Press. find
  13. Dan Hendrycks, Collin Burns, Steven Basart, Andrew Critch, Jerry Li, Dawn Song, and Jacob Steinhardt. 2021. Aligning AI With Shared Human Values. In International Conference on Learning Representations.
  14. Paul Bloom. 2010. How do morals change? Nature, 464(7288):490–490. find
  15. Kathryn Iurino and Gerard Saucier. 2020. Testing measurement invariance of the Moral Foundations Questionnaire across 27 countries. Assessment, 27(2):365–372. find
  16. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners. Advances in neural information processing systems, 33:1877–1901.
  17. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805.
  18. Danica Dillion, Niket Tandon, Yuling Gu, and Kurt Gray. 2023. Can ai language models replace human participants? Trends in Cognitive Sciences. find
  19. Denis Emelin, Ronan Le Bras, Jena D. Hwang, Maxwell Forbes, and Yejin Choi. 2021. Moral stories: Situated reasoning about norms, intents, actions, and their consequences. In Proceedings of Liwei Jiang, Jena D. Hwang, Chandrasekhar Bhagavatula, Ronan Le Bras, Maxwell Forbes, Jon Borchardt, Jenny Lian… find
  20. Kenneth Joseph and Jonathan Morgan. 2020. When do word embeddings accurately reflect surveys on our beliefs about people? In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4392–4415, Online. Association for Computational Linguistics. find
  21. Jared Kaplan, Sam McCandlish, Tom Henighan, Tom B Brown, Benjamin Chess, Rewon Child, Scott Gray, Alec Radford, Jeffrey Wu, and Dario Amodei. 2020. Scaling laws for neural language models. arXiv preprint arXiv:2001.08361.
  22. Ruibo Liu, Ge Zhang, Xinyu Feng, and Soroush Vosoughi. 2022. Aligning Generative Language Models with Human Values. In Findings of the Association for Computational Linguistics: NAACL 2022, pages 241–252. find
  23. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692. find
  24. Nicholas Lourie, Ronan Le Bras, and Yejin Choi. 2021. Scruples: A corpus of community ethical judgments on 32,000 real-life anecdotes. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 13470–13479. find
  25. Moin Nadeem, Anna Bethke, and Siva Reddy. 2021. StereoSet: Measuring stereotypical bias in pre-trained 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: Lon… find
  26. Debora Nozza, Federico Bianchi, and Dirk Hovy. 2021. HONEST: Measuring hurtful sentence completion in language models. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 2398–2406, Online. Associati…
  27. Debora Nozza, Federico Bianchi, Anne Lauscher, and Dirk Hovy. 2022. Measuring harmful sentence completion in language models for LGBTQIA+ individuals. In Proceedings of the Second Workshop on Language Technology for Equality, Diversity and Inclusion, pages 26–34, Dublin, Ireland. Association for Com… find
  28. Nedjma Ousidhoum, Xinran Zhao, Tianqing Fang, Yangqiu Song, and Dit-Yan Yeung. 2021. Probing toxic content in large pre-trained language models. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language… find
  29. Fabio Petroni, Tim Rocktäschel, Sebastian Riedel, Patrick Lewis, Anton Bakhtin, Yuxiang Wu, and Alexander Miller. 2019. Language models as knowledge bases? In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural…
  30. Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. 2019. Language models are unsupervised multitask learners. OpenAI blog, 1(8):9. find
  31. 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. Association for Computational Linguistics.
  32. PEW Research Center. 2014. Global Attitudes survey. Washington, D.C. find
  33. Maarten Sap, Saadia Gabriel, Lianhui Qin, Dan Jurafsky, Noah A. Smith, and Yejin Choi. 2020. Social bias frames: Reasoning about social and power implications of language. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 5477–5490, Online. Association…
  34. 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
  35. Zeerak Talat, Hagen Blix, Josef Valvoda, Maya Indira Ganesh, Ryan Cotterell, and Adina Williams. 2021. A Word on Machine Ethics: A Response to Jiang et al.(2021). arXiv preprint arXiv:2111.04158. — [arXiv:2111.04158](https://arxiv.org/abs/2111.04158) find
  36. Samia Touileb, Lilja Øvrelid, and Erik Velldal. 2022. Occupational biases in Norwegian and multilingual language models. In Proceedings of the 4th Workshop on Gender Bias in Natural Language Processing (GeBNLP), pages 200–211, Seattle, Washington. Association for Computational Linguistics. find
  37. Jackson Trager, Alireza S Ziabari, Aida Mostafazadeh Davani, Preni Golazazian, Farzan Karimi-Malekabadi, Ali Omrani, Zhihe Li, Brendan Kennedy, Nils Karl Reimer, Melissa Reyes, et al. 2022. The Moral Foundations Reddit Corpus. arXiv preprint arXiv:2208.05545. — [arXiv:2208.05545](https://arxiv.org/abs/2208.05545) find
  38. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention is all you need. NIPS’17, page 6000–6010, Red Hook, NY, USA. Curran Associates Inc.
  39. Jing Yi Xie, Renato Ferreira Pinto Junior, Graeme Hirst, and Yang Xu. 2019. Text-based inference of moral sentiment change. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNL… find
  40. Da Yin, Hritik Bansal, Masoud Monajatipoor, Liunian Harold Li, and Kai-Wei Chang. 2022. Geomlama: Geo-diverse commonsense probing on multilingual pre-trained language models. arXiv preprint arXiv:2205.12247. find
  41. Yukun Zhu, Ryan Kiros, Richard S. Zemel, Ruslan Salakhutdinov, Raquel Urtasun, Antonio Torralba, and Sanja Fidler. 2015. Aligning Books and Movies: Towards Story-like Visual Explanations by Watching Movies and Reading Books. CoRR, abs/1506.06724. cultural diversities and shared tendencies respective…

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/ramezani-2023-knowledge-cultural-moral-norms-large/