OmniEval
收藏资源简介:
OmniEval是由中国人民大学高瓴人工智能学院创建的一个全方位自动化的金融领域RAG评估基准数据集。该数据集包含11.4k自动生成的测试样本和1.7k人工标注的测试样本,涵盖了5个任务类别和16个金融子类别,旨在全面评估RAG系统在金融领域的性能。数据集的创建过程结合了GPT-4自动生成和人工标注,确保了数据的高质量和多样性。OmniEval的应用领域主要集中在金融领域的RAG系统评估,旨在解决RAG模型在垂直领域中的性能评估问题。
OmniEval is a comprehensive automated RAG (Retrieval-Augmented Generation) evaluation benchmark dataset for the financial domain, developed by the Gaoling School of Artificial Intelligence at Renmin University of China. This dataset comprises 11.4k automatically generated test samples and 1.7k manually annotated test samples, covering 5 task categories and 16 financial subcategories, with the core objective of comprehensively evaluating the performance of RAG systems in the financial field. The dataset's creation process integrates GPT-4-powered automatic generation and manual annotation, ensuring high data quality and diversity. The primary application scenario of OmniEval is the evaluation of RAG systems within the financial domain, aiming to address the performance evaluation challenges of RAG models in vertical specialized domains.




