CohenQu/POPE-arxiv-hard-no_guide

这是一个包含数据源、提示（内容与角色）、能力、奖励模型（地面真实与风格）、额外信息（索引与数据分割）、等级和问题ID等字段的数据集。数据集分为训练集和测试集，适用于机器学习模型的训练和评估。

This dataset includes fields such as data source, prompt (content and role), ability, reward model (ground truth and style), extra information (index and split), level, and problem ID. The dataset is divided into training and test sets, suitable for training and evaluation of machine learning models.