yguooo/summarize_from_feedback_tldr_3_filtered_oai_preprocessing_llama3_scene0

该数据集来源于OpenAI的Summarize from Feedback任务，主要用于文本摘要任务。数据集包含Reddit帖子的标题、正文、摘要等信息，并经过预处理，生成了用于摘要任务的查询和参考响应。数据集的列包括唯一标识符、子版块、标题、帖子正文、摘要、参考响应等。预处理脚本添加了查询、查询的token化版本、参考响应的token化版本等列。数据集分为训练集、验证集和测试集，分别包含116722、6447和6553个样本。

This dataset is taken from OpenAIs Summarize from Feedback task, primarily used for text summarization tasks. The dataset includes information such as titles, posts, and summaries from Reddit, and has been preprocessed to generate queries and reference responses for summarization tasks. The datasets columns include unique identifiers, subreddits, titles, post bodies, summaries, and reference responses. The preprocessing script adds columns such as queries, tokenized versions of queries, and tokenized versions of reference responses. The dataset is divided into training, validation, and test sets, containing 116722, 6447, and 6553 samples respectively.