Reliable responses from large language models (LLMs) require adherence to user instructions and retrieved information. While alignment techniques help LLMs align with human intentions and values, improving context-faithfulness through alignment remains underexplored. To address this, we propose Context-DPO, the first alignment method specifically designed to enhance LLMs' context-faithfulness. We introduce ConFiQA, a benchmark that simulates Retrieval-Augmented Generation (RAG) scenarios with knowledge conflicts to evaluate context-faithfulness. By leveraging faithful and stubborn responses to questions with provided context from ConFiQA, our Context-DPO aligns LLMs through direct preference optimization. Extensive experiments demonstrate that our Context-DPO significantly improves context-faithfulness, achieving 35% to 280% improvements on popular open-source models. Further analysis demonstrates that Context-DPO preserves LLMs' generative capabilities while providing interpretable insights into context utilization. Our code and data are released at https://github.com/byronBBL/Context-DPO
@article{arxiv.2412.15280,
title = {Context-DPO: Aligning Language Models for Context-Faithfulness},
author = {Baolong Bi and Shaohan Huang and Yiwei Wang and Tianchi Yang and Zihan Zhang and Haizhen Huang and Lingrui Mei and Junfeng Fang and Zehao Li and Furu Wei and Weiwei Deng and Feng Sun and Qi Zhang and Shenghua Liu},
journal= {arXiv preprint arXiv:2412.15280},
year = {2024}
}