Large-scale Open-domain Financial (LOFin)
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/deep-over/LOFin-bench-HiREC
下载链接
链接失效反馈官方服务：
资源简介：
该数据集名为LOFin基准，包含了145,897份美国证券交易委员会（SEC）文件和1,595组问题-答案对，旨在评估开放领域问题解答方法。数据集根据问题-答案对的格式和上下文被分为三个类别：数值型（表格）、数值型（文本）和文本型。作为一个大规模数据集，它拥有145,897份文件和1,595组问题-答案对，其任务是处理开放领域的金融问题解答。
The LOFin benchmark dataset contains 145,897 U.S. Securities and Exchange Commission (SEC) filings and 1,595 question-answer pairs, and is designed to evaluate open-domain question answering methods.
The dataset is divided into three categories based on the format and context of its question-answer pairs: numerical (tabular), numerical (textual), and textual.
As a large-scale dataset, it includes 145,897 documents and 1,595 question-answer pairs, with its core task focusing on open-domain financial question answering.
提供机构：
Authors of the paper
