C-Eval is a comprehensive Chinese evaluation suite for foundation models containing 13,948 multiple-choice questions across 52 disciplines and four difficulty levels (middle school, high school, college, and professional). It also provides a C-Eval HARD subset of especially challenging questions. The benchmark is designed to assess knowledge and reasoning abilities of Chinese/Chinese-aware large language models; the authors publish dataset files, code, and examples on the project website and GitHub, and the dataset is hosted on Hugging Face (ceval/ceval-exam). (Paper: C-Eval: A Multi-Level Multi-Discipline Chinese Evaluation Suite for Foundation Models, arXiv:2305.08322; NeurIPS 2023 Datasets & Benchmarks track.)
No results indexed yet — be the first to submit a score.
Submit a checkpoint and a reproduction script. We will run it, publish the score, and — if it takes the top — annotate the step on the progress chart with your name.