mlfoundations-dev/e1_science_ms_phi_temp2

该数据集包含了指令种子、来源、模型响应、原始行索引、长度、域名、第一个响应、推理内容、提取的解决方案、URL、文件名、成功标志、页面计数、页码、问题选项解决方案、提取的问题、提取的答案选项、匹配的解决方案、QA验证输出、分类器推理、是否为有机化学、模型标识符、最终推理轨迹、对话等字段。数据集分为训练集，包含约3.16万个示例，总大小约为6.54GB。

The dataset includes fields such as instruction seed, source, model response, original row index, length, domain, first response, reasoning content, extracted solution, URL, filename, success flag, page count, page number, question choice solutions, extracted question, extracted answer choices, matched solution, QA validation outputs, classifier reasoning, is organic chemistry, model ID, final reasoning trace, and conversations. The dataset is split into a training set with approximately 31,600 examples, totaling about 6.54GB in size.