SAIFS-AiHub/FinAR-Bench

FinAR-Bench数据集旨在评估大型语言模型在执行财务基本面分析方面的能力。数据集包含来自上海证券交易所100家上市公司2023年年度报告的财务报表。数据集分为四个部分：1) 公司表格和PDF文件，包含财务报表的PDF格式文件；2) 提取的文本，包含使用六种不同的PDF提取工具处理后的文本文件；3) 开发集，包含10家公司的评估任务；4) 测试集，包含90家公司的评估任务。每个公司的评估集包含13个任务，数据以三种格式提供：XBRL导出的Markdown表格、PDF提取的文本文件和原始PDF文件。

The FinAR-Bench dataset is designed to assess the capabilities of Large Language Models (LLMs) in performing financial fundamental analysis. The dataset contains financial statements from the 2023 annual reports of 100 listed companies on the Shanghai Stock Exchange. The dataset is divided into four parts: 1) Company Tables and PDFs, containing PDF files of financial statements; 2) Extracted Text, containing text files processed using six different PDF extraction tools; 3) Development Set, containing evaluation tasks for 10 companies; 4) Test Set, containing evaluation tasks for 90 companies. Each companys evaluation set includes 13 tasks, and the data is provided in three formats: XBRL-derived markdown tables, extracted text files from PDFs, and original PDF files.