mlfoundations-dev/pdf_qa_r1_annotated_verified

这是一个包含对话和问题解答信息的训练数据集，具体字段包括对话的发送者、消息内容、URL、页面编号、提取的答案选项、提取的解决方案、改进的问题解决方案、模型响应、原始行索引、正确性标记以及评分理由。数据集被划分为训练集，包含115328个示例，总大小为2.67GB。

This is a training dataset containing conversation and question answering information, with fields including the sender of the conversation, message content, URL, page number, extracted answer choices, extracted solution, improved question solution, model response, original row index, correctness label, and scoring reasoning. The dataset is split into a training set, containing 115328 examples, with a total size of 2.67GB.