构建全球多元价值观数据集，让大模型更贴近不同国家的真实价值。
PLURAL: A Global Dataset for Value Alignment
- 基于92国代表性调查生成50万条偏好三元组，保留文化差异信号。
- 训练后模型与目标国文化偏差降低27.7%，显著优于基线。
- 跨三国盲评验证：本地人认为其回应更符合本国价值观。
大型语言模型在全球广泛应用，却过度反映西方价值观，难以体现多元价值体系。本文提出PLURAL，一个大规模、聚焦价值观的偏好数据集，基于覆盖92个国家的整合价值观调查（IVS）。通过两阶段生成流程，将调查回答转化为合成偏好三元组，在保留规范性价值信号的同时生成真实情境。发布版本包含约50万条偏好三元组，涵盖20个多样化国家。评估显示：(i) 数据集有效保持跨国价值差异与国内多样性；(ii) 自动化评估表明，使用PLURAL训练可使模型对目标国文化特征的拟合度提升，平均绝对误差最多降低27.7%；(iii) 在印度、巴西和日本开展的176人盲评中，参与者认为基于PLURAL对齐的回复更符合本国价值观。结果表明，PLURAL具备可学习的价值引导信号，为多元化对齐提供可扩展资源。数据集：https://huggingface.co/datasets/agdhruv/plural-alignment
原文摘要 · Abstract (English)
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
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。