C-Eval: A Multi-Level Multi-Discipline Chinese Evaluation Suite for Foundation Models
This provides a new benchmark for evaluating large language models specifically for Chinese users, addressing a gap in NLP benchmarks, but it is incremental as it adapts existing evaluation concepts to a new language and context.
The authors introduced C-Eval, a comprehensive Chinese evaluation suite with multiple-choice questions across four difficulty levels and 52 disciplines, to assess foundation models' knowledge and reasoning in a Chinese context, finding that only GPT-4 achieved over 60% average accuracy, indicating significant room for improvement for current LLMs.
New NLP benchmarks are urgently needed to align with the rapid development of large language models (LLMs). We present C-Eval, the first comprehensive Chinese evaluation suite designed to assess advanced knowledge and reasoning abilities of foundation models in a Chinese context. C-Eval comprises multiple-choice questions across four difficulty levels: middle school, high school, college, and professional. The questions span 52 diverse disciplines, ranging from humanities to science and engineering. C-Eval is accompanied by C-Eval Hard, a subset of very challenging subjects in C-Eval that requires advanced reasoning abilities to solve. We conduct a comprehensive evaluation of the most advanced LLMs on C-Eval, including both English- and Chinese-oriented models. Results indicate that only GPT-4 could achieve an average accuracy of over 60%, suggesting that there is still significant room for improvement for current LLMs. We anticipate C-Eval will help analyze important strengths and shortcomings of foundation models, and foster their development and growth for Chinese users.