CohenQu/POPE-93-medium-first_guide-no_guide-easy

该数据集包含多个特征字段，包括数据来源（data_source）、提示信息（prompt，包含内容和角色）、能力（ability）、奖励模型（reward_model，包含真实值和风格）、额外信息（extra_info，包含索引和分割）、级别（level）和问题ID（problem_id）。数据集分为训练集（train）和测试集（test），分别包含384和180个样本。The dataset includes multiple feature fields such as data source (data_source), prompts (prompt, containing content and role), ability, reward model (reward_model, containing ground truth and style), extra information (extra_info, containing index and split), level, and problem ID (problem_id). The dataset is divided into training (train) and test (test) sets, containing 384 and 180 samples respectively.