Lots-of-LoRAs/task665_mmmlu_answer_generation_anatomy

task665_mmmlu_answer_generation_anatomy数据集是一个用于文本生成任务的众包数据集，包含输入和输出字符串以及唯一标识符。该数据集分为训练集、验证集和测试集，分别包含120、15和16个样本。该数据集用于支持Super-NaturalInstructions研究，该研究通过声明性指令在1600多个NLP任务上进行泛化。The task665_mmmlu_answer_generation_anatomy dataset is a crowdsourced dataset for the text-generation task, containing input and output strings along with unique identifiers. The dataset is split into training, validation, and test sets with 120, 15, and 16 examples respectively. This dataset is used to support the Super-NaturalInstructions research, which generalizes on more than 1600 NLP tasks through declarative instructions.