English

COIG-P: A High-Quality and Large-Scale Chinese Preference Dataset for Alignment with Human Values

Computation and Language 2025-04-09 v1

Abstract

Aligning large language models (LLMs) with human preferences has achieved remarkable success. However, existing Chinese preference datasets are limited by small scale, narrow domain coverage, and lack of rigorous data validation. Additionally, the reliance on human annotators for instruction and response labeling significantly constrains the scalability of human preference datasets. To address these challenges, we design an LLM-based Chinese preference dataset annotation pipeline with no human intervention. Specifically, we crawled and carefully filtered 92k high-quality Chinese queries and employed 15 mainstream LLMs to generate and score chosen-rejected response pairs. Based on it, we introduce COIG-P (Chinese Open Instruction Generalist - Preference), a high-quality, large-scale Chinese preference dataset, comprises 1,009k Chinese preference pairs spanning 6 diverse domains: Chat, Code, Math, Logic, Novel, and Role. Building upon COIG-P, to reduce the overhead of using LLMs for scoring, we trained a 8B-sized Chinese Reward Model (CRM) and meticulously constructed a Chinese Reward Benchmark (CRBench). Evaluation results based on AlignBench \citep{liu2024alignbenchbenchmarkingchinesealignment} show that that COIG-P significantly outperforms other Chinese preference datasets, and it brings significant performance improvements ranging from 2% to 12% for the Qwen2/2.5 and Infinity-Instruct-3M-0625 model series, respectively. The results on CRBench demonstrate that our CRM has a strong and robust scoring ability. We apply it to filter chosen-rejected response pairs in a test split of COIG-P, and our experiments show that it is comparable to GPT-4o in identifying low-quality samples while maintaining efficiency and cost-effectiveness. Our codes and data are released in https://github.com/multimodal-art-projection/COIG-P.

Keywords

Cite

@article{arxiv.2504.05535,
  title  = {COIG-P: A High-Quality and Large-Scale Chinese Preference Dataset for Alignment with Human Values},
  author = {P Team and Siwei Wu and Jincheng Ren and Xinrun Du and Shuyue Guo and Xingwei Qu and Yiming Liang and Jie Liu and Yunwen Li and Tianyu Zheng and Boyu Feng and Huaqing Yuan and Zenith Wang and Jiaheng Liu and Wenhao Huang and Chenglin Cai and Haoran Que and Jian Yang and Yuelin Bai and Zekun Moore Wang and Zhouliang Yu and Qunshu Lin and Ding Pan and Yuchen Jiang and Tiannan Wang and Wangchunshu Zhou and Shenzhi Wang and Xingyuan Bu and Minghao Liu and Guoyin Wang and Ge Zhang and Chenghua Lin},
  journal= {arXiv preprint arXiv:2504.05535},
  year   = {2025}
}
R2 v1 2026-06-28T22:50:08.124Z