metadata
dataset_info:
features:
- name: title
dtype: string
- name: text
dtype: string
splits:
- name: train
num_bytes: 7302223562
num_examples: 1547497
download_size: 4945898290
dataset_size: 7302223562
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
license: cc-by-sa-4.0
language:
- zh
pretty_name: Chinese Wikisource (zh-Hans, Jan 2026)
This dataset contains the main body text from Chinese Wikisource, captured in January 2026 using the site’s native zh-hans (Simplified Chinese) HTML rendering.
By relying on Wikisource’s internal conversion rules rather than third-party tools like OpenCC, the dataset avoids common errors such as region-specific vocabulary mismatches and misinterpretation of Japanese Kanji in historical texts. As a result, it provides a high-fidelity Simplified Chinese corpus of the Wikisource archive.
该数据集包含中文维基文库的正文内容，数据快照采集于 2026 年 1 月，并使用网站原生的 zh-hans（简体中文）HTML 渲染规则获取。
与通过原始数据转储并借助 OpenCC 等第三方工具转换的方式不同，本数据集依托维基文库内部的转换机制，有效避免了常见问题，如大陆与台湾地区词汇的误映射，以及历史文献中日文汉字被错误识别的情况。因此，该语料库提供了一个高保真度的简体中文维基文库文本版本。
