imminsik/summarize_from_feedback_tldr_3_filtered_oai_preprocessing_gpt2_1757492997
收藏数据链接：
官方服务：
资源简介：
这是一个从OpenAI的summarize-from-feedback任务中获取的Reddit TL;DR数据集，包含帖子的唯一标识符、来源子版块、标题、内容、摘要以及用于摘要的参考响应。经过预处理，增加了查询字段、查询和参考响应的标记化版本以及它们的长度信息。数据集分为训练集、验证集和测试集。
This is the Reddit TL;DR dataset taken from OpenAIs summarize-from-feedback task, which includes a unique identifier for the post, the subreddit it was taken from, the title, the body of the post, the summary, and the reference response for summarization. After preprocessing, it adds the query field, tokenized versions of the query and reference response, and their length information. The dataset is split into training, validation, and test sets.
提供机构：
imminsik
