Training-Free Cultural Alignment of Large Language Models via Persona Disagreement

• URL: http://arxiv.org/abs/2605.10843v2
• Date: Mon, 18 May 2026 13:29:46 GMT
• Title: Training-Free Cultural Alignment of Large Language Models via Persona Disagreement
• Abstract summary: We introduce DISCA, an inference-time method that instantiates each country as a panel of World-Values-Survey-grounded persona agents and converts their disagreement into a bounded, loss-averse logit correction.<n>Our results suggest that inference-time calibration is a scalable alternative to fine-tuning for serving the long tail of global moral preferences.
• Score: 7.507356092823674
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models increasingly mediate decisions that turn on moral judgement, yet a growing body of evidence shows that their implicit preferences are not culturally neutral. Existing cultural alignment methods either require per-country preference data and fine-tuning budgets or assume white-box access to model internals that commercial APIs do not expose. In this work, we focus on this realistic black-box, public-data-only regime and observe that within-country sociodemographic disagreement, not consensus, is the primary steering signal. We introduce DISCA (Disagreement-Informed Steering for Cultural Alignment), an inference-time method that instantiates each country as a panel of World-Values-Survey-grounded persona agents and converts their disagreement into a bounded, loss-averse logit correction. Across 20 countries and 7 open-weight backbones (2B--70B), DISCA reduces cultural misalignment on MultiTP by 10--24% on the six backbones >=3.8B, and 2--7% on open-ended scenarios, without changing any weights. Our results suggest that inference-time calibration is a scalable alternative to fine-tuning for serving the long tail of global moral preferences.

Related papers

• DiSCo: A Distribution-First Steering and Cultural Prior Evaluation Framework for Measuring Cultural Preference Bias in LLMs [1.2802720336459552]
  Large language models can systematically favour some cultures over others.<n>Existing benchmarks evaluate accuracy against a single "correct" answer.<n>We proposeo, a distribution-first forced-choice evaluation framework.
  arXiv  Detail & Related papers  (2026-09-09T14:40:08Z)
• Right Frame, Wrong Rule: Cultural Cues Expose the Financial Knowledge Gap They Were Meant to Close [57.963463622270275]
  We study normative pluralism in Islamic finance using a four-choice taxonomy.<n>A cultural cue steers a model toward one framework, but the model selects an incorrect answer within that framework.
  arXiv  Detail & Related papers  (2026-09-01T09:46:53Z)
• Whose Norms? Disentangling Cultural and Personal Alignment in Large Language Models [66.34115263472931]
  Large language models are increasingly used for social decision-making situations that require balancing cultural norms with personal preferences.<n>We introduce PACT, the Personal-Preference and Cultural-Norm Trade-off framework, which evaluates whether models choose to follow a cultural norm or allow personal preferences.<n>Our five-country human study on PACT shows that culture-following in humans is mainly driven by scenario country, with the lowest agreement when participants judge their own cultural contexts.
  arXiv  Detail & Related papers  (2026-06-05T22:20:30Z)
• Can Persona-Prompted LLMs Emulate Subgroup Values? An Empirical Analysis of Generalisability and Fairness in Cultural Alignment [75.88231994139132]
  We show that even state-of-the-art models like GPT-4.1 achieve only 57.4% accuracy in predicting modal preferences.<n>We find significant pre-existing performance biases, where models better emulate young, male, Chinese, and Christian personas.
  arXiv  Detail & Related papers  (2026-04-14T15:06:13Z)
• Does Claude's Constitution Have a Culture? [0.0]
  We evaluate Claude Sonnet on 55 World Values Survey items, selected for high cross-cultural variance across six value domains.<n>We find that Claude's value profile most closely resembles those of Northern European and Anglophone countries.<n>We discuss the compounding nature of this risk and the need for globally representative constitution-authoring processes.
  arXiv  Detail & Related papers  (2026-03-30T07:38:46Z)
• CCD-Bench: Probing Cultural Conflict in Large Language Model Decision-Making [0.9310318514564272]
  Large language models can navigate explicit conflicts between legitimately different cultural value systems.<n>CCD-Bench is a benchmark that assesses decision-making under cross-cultural value conflict.<n>CCD-Bench shifts evaluation beyond isolated bias detection toward pluralistic decision making.
  arXiv  Detail & Related papers  (2025-10-03T22:55:37Z)
• ALIGN: Word Association Learning for Cross-Cultural Generalization in Large Language Models [0.8999666725996975]
  It remains a challenge to model and align culture due to limited cultural knowledge.<n>We introduce parameter-efficient fine-tuning on native speakers' free word-association norms.<n>Our work shows that a few million culture-grounded associations can instill value alignment without costly retraining.
  arXiv  Detail & Related papers  (2025-08-19T00:55:20Z)
• The Cultural Gene of Large Language Models: A Study on the Impact of Cross-Corpus Training on Model Values and Biases [0.0]
  Large language models (LLMs) are deployed globally, yet their underlying cultural and ethical assumptions remain underexplored.<n>We compare a Western-centric model (GPT-4) and an Eastern-centric model (ERNIE Bot)<n>Human annotation shows significant and consistent divergence across both dimensions.
  arXiv  Detail & Related papers  (2025-08-17T15:54:14Z)
• CAIRe: Cultural Attribution of Images by Retrieval-Augmented Evaluation [61.130639734982395]
  We introduce CAIRe, a novel evaluation metric that assesses the degree of cultural relevance of an image.<n>Our framework grounds entities and concepts in the image to a knowledge base and uses factual information to give independent graded judgments for each culture label.
  arXiv  Detail & Related papers  (2025-06-10T17:16:23Z)
• WorldPM: Scaling Human Preference Modeling [130.23230492612214]
  We propose World Preference Modeling$ (WorldPM) to emphasize this scaling potential.<n>We collect preference data from public forums covering diverse user communities.<n>We conduct extensive training using 15M-scale data across models ranging from 1.5B to 72B parameters.
  arXiv  Detail & Related papers  (2025-05-15T17:38:37Z)
• Not All Countries Celebrate Thanksgiving: On the Cultural Dominance in Large Language Models [89.94270049334479]
  This paper identifies a cultural dominance issue within large language models (LLMs) LLMs often provide inappropriate English-culture-related answers that are not relevant to the expected culture when users ask in non-English languages.
  arXiv  Detail & Related papers  (2023-10-19T05:38:23Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.