scottgeng00/olmo-3-preference-mix-deltas_reasoning-chat35-DECON

该数据集包含了prompt、选择(chosen)和拒绝(rejected)的内容及角色信息，以及选择的模型、拒绝的模型、数据集名称、提示ID和类别等字段。数据集仅包含训练集部分，共有242106条数据，总大小约为5.86GB。提供了默认配置文件，用于指定训练数据的文件路径。The dataset includes prompt, chosen, and rejected content and role information, as well as chosen model, rejected model, dataset name, prompt ID, and category fields. The dataset contains only the training set, with a total of 242106 entries and a total size of approximately 5.86GB. A default configuration file is provided to specify the file path for the training data.