mlfoundations-dev/4o_annotated_olympiads

该数据集包含问题、推理过程、解决方案、正确性判断等信息的记录，适用于机器学习模型训练，特别是那些涉及自然语言理解和逻辑推理的任务。数据集分为训练集，共有大约34080个例子，每个例子包含了多方面的信息，如问题陈述、推理步骤、提供的解决方案、实际正确解决方案、是否正确、评估推理的理由、数据来源、对话信息等。

The dataset includes records of problems, reasoning processes, solutions, correctness judgments, and more, which is suitable for machine learning model training, especially for tasks involving natural language understanding and logical reasoning. The dataset is divided into a training set with about 34,080 examples, each containing various types of information such as problem statements, reasoning steps, proposed solutions, actual correct solutions, correctness, reasons for evaluating reasoning, data sources, and conversation information.