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

该数据集包含了用户交互的相关信息，每个交互示例包括两部分：被选中的(chosen)和被拒绝的(rejected)。每个部分包含多个特征，如内容(content)、国家(country)、语言(language)、是否涉敏(redacted)、毒性(toxic)等。此外，还有关于模型的特征，如被选中的模型(chosen_model)和被拒绝的模型(rejected_model)。数据集分为训练集(train)，提供了训练集的数据大小和示例数量。数据集适用于自然语言处理任务，如文本分类、情感分析等。

The dataset consists of user interaction information, with each interaction example including two parts: chosen and rejected. Each part contains multiple features such as content, country, language, whether sensitive (redacted), toxicity (toxic), etc. In addition, there are features about the models, such as the chosen model (chosen_model) and the rejected model (rejected_model). The dataset is split into a training set (train), and the data size and number of examples in the training set are provided. The dataset is suitable for natural language processing tasks such as text classification, sentiment analysis, etc.