VGraf/olmo-2-0325-32b-preference-mix

该数据集包含了选中的(chosen)和被拒绝的(rejected)信息，每个信息包括内容和角色。还包括了选中模型的名称(chosen_model)和拒绝模型的名称(rejected_model)，以及唯一标识符(id)和数据来源(source)。此外，还包含了一系列消息(messages)，每条消息也包括内容和角色。数据集分为训练集(train)，共有377,809个示例，大小约为3.5GB。

The dataset includes selected (chosen) and rejected (rejected) information, each containing content and role. It also includes the name of the chosen model (chosen_model) and the name of the rejected model (rejected_model), a unique identifier (id), and the data source (source). Additionally, it contains a series of messages (messages), with each message including content and role. The dataset is split into a training set (train) with a total of 377,809 examples, approximately 3.5GB in size.