dworsleytonks/project14-llm-original-dataset

该数据集包含多个配置，专注于医学领域的问答和偏好任务。具体配置包括：1) frenchmedmcqa：包含医学多选题，问题、答案选项、正确答案、科目名称等特征；2) mediqal_mcqm和mediqal_mcqu：医学多选题数据集，包含临床案例、问题、答案选项、正确答案等；3) mediqal_oeq：医学开放式问题数据集，包含临床案例、问题、答案等；4) medquad：医学问答数据集，包含问题类型、问题和答案；5) ultramedical_preference：医学偏好数据集，包含提示、选择答案、拒绝答案、元数据和反馈等。每个配置都有训练、验证和测试集（部分配置仅有测试集），并详细记录了每个分区的字节数和示例数。

The dataset comprises multiple configurations focused on medical question-answering and preference tasks. Specific configurations include: 1) frenchmedmcqa: Contains medical multiple-choice questions with features like questions, answer options, correct answers, subject names, etc.; 2) mediqal_mcqm and mediqal_mcqu: Medical multiple-choice question datasets with clinical cases, questions, answer options, correct answers, etc.; 3) mediqal_oeq: Medical open-ended question dataset with clinical cases, questions, answers, etc.; 4) medquad: Medical Q&A dataset with question types, questions, and answers; 5) ultramedical_preference: Medical preference dataset with prompts, chosen answers, rejected answers, metadata, and feedback. Each configuration has train, validation, and test splits (some only have test splits), with detailed records of bytes and examples per split.