DiverValue-Bench: A Benchmark and Fine-Tuning Framework for Aligning Large Language Models with Diverse Human Values

Liang, Yao; Zhao, Dongcheng; Zhao, Feifei; Shen, Guobin; Wang, Yuwei; Liang, Dongqi; Zeng, Yi

doi:10.24963/ijcai.2026/850

DiverValue-Bench: A Benchmark and Fine-Tuning Framework for Aligning Large Language Models with Diverse Human Values

Yao Liang, Dongcheng Zhao, Feifei Zhao, Guobin Shen, Yuwei Wang, Dongqi Liang, Yi Zeng

Download BibTeX Cite

Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence. Human-Centred AI. Pages 7647-7657.

https://doi.org/10.24963/ijcai.2026/850

Keywords

Human-Centred AI: Humans and AI
Human-Centred AI: Natural Language Processing
Human-Centred AI: AI Ethics, Trust, Fairnes

Abstract

Aligning large language models (LLMs) with diverse human values is essential for safe and effective deployment, yet existing benchmarks often overlook cultural and demographic variation. We introduce DiverValue-Bench, a population-aware benchmark for evaluating multi-dimensional value alignment across 74 countries/regions. It contains 23,763 quality-controlled instances derived from PRISM user feedback and audited through large-scale human validation, with fine-grained value labels, personalized questions, contrastive reference answers, and rich demographic metadata. Using DiverValue-Bench, we evaluate representative LLMs and reveal substantial geographic and demographic disparities that are masked by aggregate performance. We further show that lightweight preference-based fine-tuning with Low-Rank Adaptation (LoRA) and Direct Preference Optimization (DPO) substantially improves in-domain value alignment while yielding consistent out-of-domain gains. These results highlight the need for population-aware alignment evaluation and demonstrate the utility of DiverValue-Bench as a practical foundation for global alignment, personalized value modeling, and equitable AI development.

Yao Liang, Dongcheng Zhao, Feifei Zhao, Guobin Shen, Yuwei Wang, Dongqi Liang, and Yi Zeng. 2026. DiverValue-Bench: A Benchmark and Fine-Tuning Framework for Aligning Large Language Models with Diverse Human Values. In Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26. IJCAI, 7647–7657. https://doi.org/10.24963/ijcai.2026/850

@inproceedings{ijcai2026p850,
  title     = {DiverValue-Bench: A Benchmark and Fine-Tuning Framework for Aligning Large Language Models with Diverse Human Values},
  author    = {Liang, Yao and Zhao, Dongcheng and Zhao, Feifei and Shen, Guobin and Wang, Yuwei and Liang, Dongqi and Zeng, Yi},
  booktitle = {Proceedings of the Thirty-Fifth International Joint Conference on
               Artificial Intelligence, {IJCAI-26}},
  publisher = {International Joint Conferences on Artificial Intelligence Organization},
  editor    = {Diego Calvanese},
  pages     = {7647--7657},
  year      = {2026},
  month     = {8},
  note      = {Human-Centred AI},
  doi       = {10.24963/ijcai.2026/850},
  url       = {https://doi.org/10.24963/ijcai.2026/850},
}

TY  - CPAPER
TI  - DiverValue-Bench: A Benchmark and Fine-Tuning Framework for Aligning Large Language Models with Diverse Human Values
AU  - Liang, Yao
AU  - Zhao, Dongcheng
AU  - Zhao, Feifei
AU  - Shen, Guobin
AU  - Wang, Yuwei
AU  - Liang, Dongqi
AU  - Zeng, Yi
T2  - Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26
PB  - International Joint Conferences on Artificial Intelligence Organization
PY  - 2026
SP  - 7647
EP  - 7657
DO  - 10.24963/ijcai.2026/850
UR  - https://doi.org/10.24963/ijcai.2026/850
ER  -