scottgeng00/olmo-3-preference-mix-deltas-yolo_downsample_wildchat_to_multilingual

这是一个包含用户交互信息的文本数据集，包括用户的语言、国家、IP地址等信息，以及与交互相关的文本内容。数据集分为被选中(chosen)和被拒绝(rejected)两部分，每一部分都包含多个字段，用于记录不同的交互特征。此外，数据集还记录了模型选择、数据集名称、prompt ID和分类信息。训练集包含约1755084个示例，总数据大小约为14.53GB。

This is a text dataset containing user interaction information, including users language, country, IP address, etc., and text content related to interactions. The dataset is divided into two parts: chosen and rejected, each containing multiple fields to record different interaction features. In addition, the dataset also records model selection, dataset name, prompt ID, and category information. The training set contains about 1,755,084 examples, with a total data size of approximately 14.53GB.