saepark/ultrafeedback-binarized-preferences-medical-cldfilter-train

这是一个包含文本数据的训练集，其中每个样本由源文本(source)、提示(prompt)、选中的文本内容(chosen.content)、选中角色的类型(chosen.role)、选中文本的评分(chosen-rating)、选中文本的模型(chosen-model)、拒绝的文本内容(rejected.content)、拒绝角色的类型(rejected.role)、拒绝文本的评分(rejected-rating)和拒绝文本的模型(rejected-model)组成。数据集分为训练集(train)部分，大小为约241MB。

This is a training dataset containing text data, where each sample consists of source text (source), prompt (prompt), selected text content (chosen.content), type of selected role (chosen.role), rating of selected text (chosen-rating), model of selected text (chosen-model), rejected text content (rejected.content), type of rejected role (rejected.role), rating of rejected text (rejected-rating), and model of rejected text (rejected-model). The dataset is split into a training set (train) part, with a size of approximately 241MB.