Compiling a Danish Benchmark Dataset for Assessing Selected Reasoning Capabilities of Large Language Models

The project compiles a robust general reasoning benchmark dataset for Danish for assessing large language models (LLMs). Our dataset derives from theory about: 1) how our language and vocabulary work as a system, 2) how humans think and act with language, and 3) how the Danish language community differs from e.g. English wrt. linguistic expressivity and cultural identity.