Instruction-tuning Dataset

该数据集旨在为开源大型语言模型进行指令调优而创建，包含了经过非单调逻辑推理模型验证的训练和验证样本。这些样本经过筛选，确保了正确的验证标签，用于对开源大型语言模型进行微调。该数据集的规模包括1700个训练样本和200个验证样本，其任务专注于生物医学领域的自然语言推理指令调优。

This dataset is constructed for instruction tuning of open-source large language models, and encompasses training and validation samples verified by non-monotonic logical reasoning models. All samples have been screened to ensure accurate validation labels, which are dedicated to fine-tuning open-source large language models. The dataset comprises 1700 training samples and 200 validation samples, with its tasks centered on natural language reasoning instruction tuning within the biomedical domain.