InsQABench

该数据集是为中国保险行业设立的一个基准数据集，它被分为三个类别：保险常识知识、保险结构化数据库以及保险非结构化文档。该数据集包含了从用户提问和专家输入中精心筛选的问答对，确保了在现实世界场景中的高质量和代表性。其规模包括1万个训练样本和990个测试样本，任务集中在保险领域的问答。This dataset is a benchmark for the Chinese insurance industry, divided into three categories: insurance common-sense knowledge, insurance structured databases, and insurance unstructured documents. It contains carefully curated question-answer pairs sourced from user queries and expert inputs, ensuring high quality and real-world representativeness. The dataset comprises 10,000 training samples and 990 test samples, with tasks focused on question answering in the insurance domain.

搜集汇总

背景与挑战

背景概述

InsQABench是一个专为中国保险行业设计的问答基准数据集，分为保险常识知识、结构化数据库和非结构化文档三个类别，包含从用户和专家输入中筛选的高质量问答对。该数据集规模为1万个训练样本和990个测试样本，专注于支持保险领域的现实世界问答任务。