LLMEval

LLMEval is an academic initiative from Fudan NLP Lab, providing comprehensive and rigorous evaluation frameworks for large language models (LLMs). It offers curated, contamination-resistant benchmarks covering over 13 academic disciplines, medical AI, and logical reasoning, with a database of 220,000+ questions and a robust longitudinal evaluation pipeline. The platform is designed for AI researchers, practitioners, and developers seeking accurate, robust, and fair performance analysis of LLMs.