价格
免费
下载次数
0
点赞数
0
A-Eval is a benchmark designed to evaluate Chat LLMs of various scales from a practical application perspective. The dataset includes 678 question-and-answer pairs spanning 5 categories, 27 sub-categories, and 3 difficulty levels. A-Eval offers clear empirical and engineering guidelines for choosing the “best” model for real-world applications.
| 来源类型 | API 接入 |
| 开源状态 | ✅ 开源 |
| 授权协议 | apache-2.0 |
| 数据语言 | - |
| 数据量级 | - |
| 交付格式 | - |
| 仓库路径 | cubeAI-yuanjing/A-Eval |
| 公开状态 | 公开 |
该数据集为开源数据集，可通过 Git 方式免费下载
git lfs install && git clone https://aihub.caict.ac.cn/datasets/cubeAI-yuanjing/A-Eval.git