purewhite42/CoPA_Dataset

Chain-of-Proxy-Autoformalization数据集集合用于训练在形式数学和Lean 4相关的各种任务中的模型。数据集包括不同的分割，用于任务，如声明自动形式化，解决方案草案，证明和解决方案步骤预测和生成。每个任务都有自己的训练数据集，组织在JSONL文件中，包含系统提示和用户输入/输出之间的对话。数据集描述详细，包括其用法，性能指标，以及每个任务的示例。README还包括引用信息和许可详情。

The Chain-of-Proxy-Autoformalization Dataset Collection is used for training models in various tasks related to formal mathematics and Lean 4. The dataset includes different splits for tasks such as statement autoformalization, solution drafting, proof and solution step prediction and generation. Each task has its own set of training data organized in JSONL files containing conversations between a system prompt and user input/output. The dataset is described in detail, including its usage, performance metrics, and examples for each task. The README also includes citation information and licensing details.