Benchmarks in Leipzig

A group of mathematicians created a dataset of research-level mathematics questions with known answers and evaluated the performance of large language models (LLMs) in solving these questions.