Sign in to view source links and access this dataset
Description
Taiwan's legal system is the focus of this 1,652-pair traditional Chinese question-answer dataset covering 993 distinct statutes. Created by lianghsun, it provides article content, legal background, and practical explanations, formatted for both supervised fine-tuning and direct preference optimization. The dataset was last updated on April 12, 2026.
Use Cases
Supervised fine-tuning of legal chatbots based on structured article explanations
Direct preference optimization for aligning model outputs with high-quality legal responses
Benchmarking model performance on retrieving and explaining Taiwan's legal statutes
Training models to generate structured legal text with background context
Strengths
Contains 1,652 question-answer pairs for model training
Covers 993 different statutes, indicating broad legal coverage
Includes paired chosen/rejected responses specifically formatted for DPO training
Provides two processed subsets (default and processed) for different training workflows
Limitations
Column-level documentation is absent; field semantics must be inferred after download
Row count is unknown, which may limit suitability assessment
Data may reflect geographic bias inherent to its focus on Taiwan's legal system
Provenance
Source
huggingface
Collection Method
Likely compiled from the statutes of the Republic of China (Taiwan).
Freshness
Last updated 2026-04-12 10:31:37; freshness should be verified
Geography
Taiwan
License is unknown and should be verified before use.