Search papers, labs, and topics across Lattice.
This paper introduces PLURAL, a comprehensive dataset designed to enhance the value alignment of large language models (LLMs) by incorporating diverse cultural perspectives from 92 countries. Utilizing a two-stage generation pipeline, the authors convert survey responses from the Integrated Values Survey into approximately 500,000 synthetic preference triplets that reflect normative values across various nations. Evaluation results indicate that training on PLURAL significantly improves alignment with cultural profiles, achieving a reduction in mean absolute error by up to 27.7% compared to existing baselines, and human evaluators affirm its effectiveness in representing national values.
Training LLMs on the PLURAL dataset can reduce cultural misalignment by nearly 28%, making them more representative of global values.
Large language models (LLMs) are used worldwide, yet disproportionately reflect Western values, limiting their ability to represent diverse value systems. We introduce PLURAL, a large-scale, value-focused preference dataset grounded in the Integrated Values Survey (IVS), a nationally representative survey spanning 92 countries. Using a two-stage generation pipeline, we transform survey responses into synthetic preference triplets that preserve normative value signals while producing realistic scenarios. We release an initial version of PLURAL containing ~500,000 preference triplets representing people in 20 diverse countries. We evaluate PLURAL in three ways: (i) dataset-level validation showing that it preserves both cross-country value differences and within-country diversity from the original survey; (ii) automated evaluation showing that training on PLURAL improves alignment with target countries'cultural profiles, reducing mean absolute error by up to 27.7% relative to strong baselines; and (iii) blind human evaluation with 176 evaluators in India, Brazil, and Japan, who judge PLURAL-aligned responses as more representative of their national values. Together, these results show that PLURAL contains learnable signal for value steering, offering a scalable resource for pluralistic alignment. Dataset: https://huggingface.co/datasets/agdhruv/plural-alignment
