Carefully Considering Culture: Analyzing LLM Alignment in Single- and Multi-Cultural Settings using Cultural Consensus Theory
On the map: LLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Abstract (arXiv)
Recent work in NLP has probed large language models for their understanding of cultural norms across countries. However, this work typically considers distributional patterns, ignoring group consensus or possible multicultural environments within a country. In this work, we leverage cultural consensus theory (CCT) from cultural anthropology to model such multidimensional nuance. Applying CCT to the World Values Survey (WVS) across 10 countries and 12 domains, we demonstrate that models frequently misrepresent cultural structures by either failing to form cohesive consensus or severely over-regularizing consensus. Through explicit representation of intra-group variance, CCT provides actionable diagnostics to evaluate when models reflect true human diversity versus algorithmic homogenization.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Human values and AI feedback alignment, one of 10 directions in LLM alignment / RLHF / preference optimisation. Builds on 11 of its 29 references that are in the map. First posted on arXiv 29 May 2026.
Builds on
11 of 29 references.
Closest ancestors · 1
Same direction, newest first
Other references · 10
- 2023Whose Opinions Do Language Models Reflect?cited by 773
- 2023Towards Measuring the Representation of Subjective Global Opinions in Language Modelscited by 437
- 2024Global MMLU: Understanding and Addressing Cultural and Linguistic Biases in Multilingual Evaluationcited by 191
- 2024Political Compass or Spinning Arrow? Towards More Meaningful Evaluations for Values and Opinions in Large Language Modelscited by 162
- 2024Survey of Cultural Awareness in Language Models: Text and Beyondcited by 103
- 2018The Geometry of Culture: Analyzing Meaning through Word Embeddingscited by 87
- 2025Randomness, Not Representation: The Unreliability of Evaluating Cultural Alignment in LLMscited by 54
- 2025Few-shot LLM Synthetic Data with Distribution Matchingcited by 10
- 2025Open WebUI: An Open, Extensible, and Usable Interface for AI Interactioncited by 6
- 2022An Open-Source Cultural Consensus Approach to Name-Based Gender Classificationcited by 5
Built on by
Not cited yet.
Similar, not cited
Close in wording, no citation either way
- 2026Aligned but Flattened: Analyzing the Trade-off between Cultural Alignment and Diversity in LLMs
- 2025Break the Checkbox: Challenging Closed-Style Evaluations of Cultural Alignment in LLMs
- 2023Cultural Alignment in Large Language Models: An Explanatory Analysis Based on Hofstede's Cultural Dimensions
- 2025Multilingual != Multicultural: Evaluating Gaps Between Multilingual Capabilities and Cultural Alignment in LLMs
- 2026Whose Norms? Disentangling Cultural and Personal Alignment in Large Language Models
- 2026AlignCultura: Towards Culturally Aligned Large Language Models?
- 2026Understanding Cultural Alignment in Multilingual LLMs via Natural Debate Statements ↗
- 2026Cultural Value Alignment Via Latent Activation Steering in Large Language Models