ChineseWebText 2.0: Large-Scale High-quality Chinese Web Text with Multi-dimensional and fine-grained information

TL;DR

ChineseWebText 2.0 is a large-scale, high-quality Chinese dataset with multi-dimensional fine-grained annotations, designed to improve LLM training by providing detailed quality, domain, and toxicity information.

Contribution

This paper introduces MDFG-tool, a novel data construction pipeline that creates a comprehensive Chinese text dataset with detailed annotations, surpassing previous datasets in scale and granularity.

Findings

Releases 3.8TB of ChineseWebText 2.0 data

Provides fine-grained labels including quality, domain, and toxicity

Facilitates targeted data selection for LLM training

Abstract

During the development of large language models (LLMs), pre-training data play a critical role in shaping LLMs' capabilities. In recent years several large-scale and high-quality pre-training datasets have been released to accelerate the research of LLMs, including ChineseWebText1.0, C4, Pile, WanJuan, MAPCC and others. However, as LLMs continue to evolve, focus has increasingly shifted to domain-specific capabilities and safety concerns, making those previous coarse-grained texts insufficient for meeting training requirements. Furthermore, fine-grained information, such as quality, domain and toxicity, is becoming increasingly important in building powerful and reliable LLMs for various scenarios. To address these challenges, in this paper we propose a new tool-chain called MDFG-tool for constructing large-scale and high-quality Chinese datasets with multi-dimensional and fine-grained information. First, we employ manually crafted rules to discard explicit noisy texts from raw contents. Second, the quality evaluation model, domain classifier, and toxicity evaluation model are well-designed to assess the remaining cleaned data respectively. Finally, we integrate these three types of fine-grained information for each text. With this approach, we release the largest, high-quality and fine-grained Chinese text ChineseWebText2.0, which consists of 3.8TB and each text is associated with a quality score, domain labels, a toxicity label and a toxicity score, facilitating the LLM researchers to select data based on various types of fine-grained information. The data, codes and the tool-chain are available on this website https://github.com/CASIA-LM/ChineseWebText-2.0