redue/4chan-10k-shuffled

该数据集是adamo1139/4chan_archive_ShareGPT_only5score_speakerCount_only2speakers数据集的子集，经过缩减至10,000行并打乱顺序，以保持原始数据中的板块多样性。数据集包含多个特征，如id、conversations（包含发言者和内容）、comments、scores_list、final_score、speaker_comments和speaker_count，主要用于训练目的，具体应用于自然语言处理任务，如对话生成或评分分析。

Subset of adamo1139/4chan_archive_ShareGPT_only5score_speakerCount_only2speakers, reduced to 10K rows and shuffled to maintain board variety. The dataset includes features such as id, conversations (with from and value fields), comments, scores_list, final_score, speaker_comments, and speaker_count, intended for training purposes in NLP tasks like dialogue generation or scoring analysis.