ndhananj/uplimit-synthetic-data-week-2-filtered-basic

偏好调整数据集用于Uplimit合成数据课程第二周。数据集包含多个字段，如系统信息、输入、选择的答案、拒绝的答案、生成的内容、顺序、标注模型、标注提示、原始标注响应、评分、评分依据、状态、原始选择的答案、原始拒绝的答案、选择的得分、是否在GSM8k训练集中、可读性分数和领域。数据集分为训练集，其大小为19070396字节，包含1468个示例。

Preference Tuning Dataset for Uplimit Synthetic Data Course Week2. The dataset includes fields such as system information, input, chosen answer, rejected answer, generated content, order, labeling model, labeling prompt, raw labeling response, rating, rationale, status, original chosen answer, original rejected answer, chosen score, whether in GSM8k training set, readability score, and domain. The dataset is split into a training set, which is 19070396 bytes in size and contains 1468 examples.