quanda-bench-test/trex-subset

该数据集包含四个字段：prompt（提示）、answer（答案）、alt_answers（备选答案）和evidence_sentences（证据句子）。数据集专为训练机器学习模型而设计，提供了训练集。每个示例由一个提示和相应的答案、备选答案以及支持答案的证据句子组成。

The dataset includes four fields: prompt, answer, alt_answers, and evidence_sentences. It is designed for training machine learning models and provides a training set. Each example consists of a prompt along with the corresponding answer, alternative answers, and evidence sentences supporting the answer.