dogtooth/uf_tulu_2_uf_iter2_small_beta_2

该数据集包含一个提示(prompt)字段，一系列消息(messages)字段，以及模型完成(model_completion)和参考完成(reference_completion)字段。消息字段进一步包含内容(content)和角色(role)信息。数据集分为训练集(train)，共有约122270个样本，总数据大小约为703MB。

The dataset includes a prompt field, a series of messages fields, and model completion and reference completion fields. The messages field further contains content and role information. The dataset is split into a training set (train) with approximately 122270 samples and a total data size of about 703MB.