SKIML-ICL/temp_nq

该数据集包含多个字段，涵盖了上下文、问题、答案、实体类型等信息。数据集被分为训练集和验证集，训练集包含20060个样本，验证集包含1508个样本。数据集的字段包括子集、上下文、上下文标记、问题ID、问题、问题标记、检测到的答案、答案、答案句子、实体类型、相似实体、相似实体得分、随机实体、随机实体得分、冲突句子、冲突上下文、上下文列表、是否有答案、可回答性、掩码查询、掩码查询嵌入、问答案例、不可回答案例、冲突案例、恶魔和文本等。

The dataset contains multiple fields, including context, questions, answers, entity types, and more. The dataset is divided into a training set and a validation set, with the training set containing 20,060 samples and the validation set containing 1,508 samples. The fields in the dataset include subset, context, context tokens, question ID, question, question tokens, detected answers, answers, answer sentence, entity type, similar entity, similar entity score, random entity, random entity score, conflict sentence, conflict context, context list, has answer, answerable, masked query, masked query embedding, QA cases, unanswerable cases, conflict cases, demons, and text.