SJTULean/LeanStatement_CoT

LeanStatement_CoT是一个专门设计的数据集，用于训练模型将自然语言中的数学定理翻译为Lean4的正式声明，并包含链式推理（CoT）步骤。数据集大约包含142,000行数据，采用JSON格式。该数据集的创建者是SJTULean，并遵循Apache-2.0许可证。LeanStatement_CoT is a specialized dataset designed for training models to translate mathematical theorems from natural language into Lean4 formal statements, including Chain-of-Thought (CoT) reasoning steps. The dataset contains approximately 142,000 entries in JSON format. It was created by SJTULean and is licensed under Apache-2.0.