snap-stanford/outputs_final_rl_r_only_gpt5_judge_train_225_persona_interests_post_dist

该数据集包含多个字段，主要涉及对话或文本生成任务。具体包括：数据集索引（dataset_index）、响应键（response_key）、提示信息（prompt，包含内容、名称和角色）、生成文本（generation）、额外信息（extra_info，包含评论指标、层次名称、索引、是否顶级、媒体来源、名称、人物设定、帖子、帖子ID、帖子指标、原始提示和分割信息）以及多种指标（metrics，涵盖内容类型、情感、敌意、幽默、情绪和立场等）。数据集分为训练集，包含3366个示例，总大小约为88MB。The dataset includes multiple fields primarily related to dialogue or text generation tasks. Specifically, it contains: dataset index (dataset_index), response key (response_key), prompt information (prompt, including content, name, and role), generated text (generation), extra information (extra_info, including comment metrics, hierarchy name, index, is_top_level, media source, name, persona, post, post_id, post metrics, raw prompt, and split), and various metrics (metrics, covering content type, emotion, hostility, humor, sentiment, and stance). The dataset is divided into a training set with 3,366 examples and a total size of approximately 88MB.