semi_qwen2.5_0.5B

该数据集包含一个训练集，共4957个样本，总大小为7195704字节。数据集的结构包括多个字段：'id'（字符串类型）、'question_stem'（问题题干，字符串类型）、'choices'（选项列表，包含文本和标签）、'answerKey'（正确答案键，字符串类型）、'fact1'（事实描述，字符串类型）以及多个与事实和答案相关的权重和逻辑标签字段（如'W_base'、'fact_h'、'answer_h'等）。数据集适用于需要处理问题回答、事实推理及相关权重计算的机器学习任务。

This dataset contains a training set comprising 4957 samples with a total size of 7195704 bytes. The dataset structure includes multiple fields: 'id' (string type), 'question_stem' (question stem, string type), 'choices' (list of options containing text and labels), 'answerKey' (correct answer key, string type), 'fact1' (fact description, string type), as well as several weight and logic label fields related to facts and answers (such as 'W_base', 'fact_h', 'answer_h', etc.). This dataset is applicable to machine learning tasks that require handling question answering, factual reasoning and related weight calculation.