English

Leveraging Web-Crawled Data for High-Quality Fine-Tuning

Computation and Language 2024-08-16 v1

Abstract

Most large language models are fine-tuned using either expensive human-annotated data or GPT-4 generated data which cannot guarantee performance in certain domains. We argue that although the web-crawled data often has formatting errors causing semantic inaccuracies, it can still serve as a valuable source for high-quality supervised fine-tuning in specific domains without relying on advanced models like GPT-4. To this end, we create a paired training dataset automatically by aligning web-crawled data with a smaller set of high-quality data. By training a language model on this dataset, we can convert web data with irregular formats into high-quality ones. Our experiments show that training with the model-transformed data yields better results, surpassing training with only high-quality data by an average score of 9.4% in Chinese math problems. Additionally, our 7B model outperforms several open-source models larger than 32B and surpasses well-known closed-source models such as GPT-3.5, highlighting the efficacy of our approach.

Keywords

Cite

@article{arxiv.2408.08003,
  title  = {Leveraging Web-Crawled Data for High-Quality Fine-Tuning},
  author = {Jing Zhou and Chenglin Jiang and Wei Shen and Xiao Zhou and Xiaonan He},
  journal= {arXiv preprint arXiv:2408.08003},
  year   = {2024}
}
R2 v1 2026-06-28T18:13:32.660Z