sapienzanlp/zebra-kb-explanations

ZEBRA数据集是一个用于常识问答（CQA）的零样本学习数据集，专为大型语言模型（LLM）设计。它包括不同的配置，每个配置都有自己的特征和分割。数据集的结构旨在通过示例检索、知识生成和有信息推理阶段来促进零样本学习。README还包括安装说明、快速入门指南、模型和数据可用性、性能指标、引用信息和许可详情。

ZEBRA dataset is a zero-shot learning dataset designed for Commonsense Question Answering (CQA), specifically tailored for large language models (LLMs). It includes various configurations, each with its own features and splits. The dataset is structured to facilitate zero-shot learning through stages of example retrieval, knowledge generation, and informed reasoning. The README also provides installation instructions, a quick start guide, information on model and data availability, performance metrics, citation details, and licensing information.