scottgeng00/olmo-3-preference-mix-deltas_reasoning-chosen_qwen8b-chat35-DECON

该数据集包含了提示(prompt)、选中(chosen)和拒绝(rejected)的内容及角色信息，以及选中模型(chosen_model)和拒绝模型(rejected_model)的相关信息。数据集被划分为训练集(train)，共有220,840个示例，大小为约4.84GB。此外，提供了默认配置(default)，指定了训练数据文件的路径。The dataset includes prompt, chosen, and rejected content and role information, as well as information about the chosen model (chosen_model) and the rejected model (rejected_model). The dataset is split into a training set (train) with a total of 220,840 examples and a size of approximately 4.84GB. In addition, a default configuration (default) is provided, specifying the path to the training data files.