18,300+ question-answer pairs derived from Turkish legal texts, case law, and legislation. The collection focuses on the Turkish judicial system, providing structured data for domain-specific natural language processing tasks in the Turkish language.
Use Cases
- Fine-tune Large Language Models (LLMs) for the legal domain using the question and answer pairs.
- Benchmark Retrieval-Augmented Generation (RAG) systems by comparing generated responses against the ground-truth answer column.
- Develop automated legal advisory tools by training on the question-answer pairs derived from Turkish legislation.
Strengths
- 18,300+ question-answer pairs covering diverse aspects of Turkish law.
- Data compiled from official sources including case law (içtihat) and legislation (mevzuat).
- Structured specifically for fine-tuning Large Language Models (LLMs) and RAG system evaluation.