Structured Taiwanese Law Articles for SFT, Hundreds of Thousands of Records

Taiwan's tw-structured-law-article is a traditional Chinese instruction-tuning dataset containing hundreds of thousands of structured legal articles from the Republic of China. It was created by lianghsun and focuses on preserving diverse formatting like hierarchical clauses, tables, and attachments. The dataset was last updated on 2026-04-15.

Use Cases

• Fine-tuning LLMs for legal text generation based on structured article formats described.
• Training models to understand hierarchical legal clauses (articles, items, sub-items) as mentioned.
• Improving model comprehension of legal documents containing tables, appendices, and attachments.
• Creating benchmarks for evaluating LLMs on structured legal text parsing and reasoning.

Strengths

• Contains hundreds of thousands of records, as indicated by the description.
• Explicitly focuses on preserving diverse legal formatting (hierarchical clauses, tables, attachments).
• Processes content from PDF attachments to capture a broader range of legal text.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last updated 2026-04-15; freshness should be verified.

Provenance

Source

Author lianghsun on Hugging Face.

Collection Method

Likely processed from legal documents of the Republic of China (Taiwan).

Freshness

Last updated 2026-04-15 02:01:53.

Geography

Republic of China (Taiwan).