Skip to content

Dataset · lianghsun

Taiwan legal corpus in Traditional Chinese

lianghsun/tw-legal-oa-corpus

Over 33 thousand open legal articles and publications in Traditional Chinese for fine-tuning language models.

What's inside

Contains full texts of free legal journals and publications of Taiwan in Traditional Chinese, split into subsets for each publication. The corpus has 33,358 documents (about 521.7 million tokens using cl100k_base), collected from the open resources list of the Judicial Reform Foundation. The 19-field schema completely matches the fineweb-edu-zhtw dataset, allowing easy combination during training.

What you can do with it

  • Continued pretraining of base LLMs in Traditional Chinese
  • Adapting language models to legal domains and Taiwan law
  • Creating specialized legal search and reference systems
Tasks
Text generation
Languages
Chinese
Size
10K–100K rows
Downloads this month
35.8K
Likes
0
Access
on request on Hugging Face

How to load it

from datasets import load_dataset

ds = load_dataset("lianghsun/tw-legal-oa-corpus")