Dataset for Evaluating LLMs in Thematic Analysis of Czech History Curricula

This comprehensive dataset includes a collection of Czech school curricula, sample data, processed curricula, and the results of thematic analysis tasks. It was developed to assess the performance of large language models (LLMs) in tasks such as data extraction, topic splitting, and topic labelling. The dataset features curricula in various formats (PDF, HTML, DOCX) standardized to PDF, along with processed outputs and analysis results. This resource is valuable for researchers exploring the application of LLMs in educational data analysis and supports comparative studies of curriculum content across different educational contexts.