mlfoundations-dev/pdf_sci_questions_difficulty_filter_and_science_question_labeler

这是一个包含网页问题和答案的数据集，其中包括网页URL、页面编号、提取的答案选项、提取的解决方案、改进后的问答解决方案、模型响应、原始行索引、问题分类及其推理、难度及其推理等字段。训练集包含10000个样本，数据集大小为173,676,983字节。

This is a dataset containing web-based questions and answers, which includes fields such as URL, page number, extracted answer choices, extracted solution, improved question solution, model response, original row index, question classification and reasoning, difficulty and reasoning. The training set contains 10,000 samples, and the dataset size is 173,676,983 bytes.