RAGEVALUATION-HJKMY/ragbench_techqa_400row

该数据集包含多个特征字段，如问题(question)、文档(documents)、回答(response)等。每个样本都有唯一的id标识。文档和回答都被进一步细分为句子。此外，数据集还包含关于回答生成和标注所使用的模型名称，以及各种评分指标，如遵守度评分(adherence_score)、相关性解释(relevance_explanation)等。数据集分为训练集(train)，共有400个样本，总大小约为18MB。

The dataset includes multiple feature fields such as question, documents, response, etc. Each sample has a unique id. Both documents and responses are further subdivided into sentences. Additionally, the dataset contains the names of the models used for response generation and annotation, as well as various scoring metrics such as adherence score, relevance explanation, etc. The dataset is split into a training set (train) with a total of 400 samples, with an overall size of approximately 18MB.