lcalvobartolome/rosie_mind_topics

ROSIE-MIND-Topics是一个包含875,230段落的英语-西班牙语双语数据集，其中的主题模型信息是通过在这份数据上训练PLTM模型并提取30个主题得到的。该数据集作为MIND管道的输入数据集，用于检测问答对中的多语言和文化差异。每条记录包括段落和对应的完整文档、预处理输出（词形还原、翻译）以及主题模型特征（主题分布和主要主题）。ROSIE-MIND-Topics is a bilingual (English-Spanish) dataset of 875,230 passages containing topic modeling information derived from training a PLTM model on this data with 30 topics. It serves as the input dataset for the MIND pipeline, which detects multilingual and cultural discrepancies in question-answer pairs. Each record includes the passage and corresponding full document, preprocessing outputs (lemmas, translations), and topic model features (topic distributions and dominant topic).