Large language models have demonstrated outstanding performance in various natural language processing tasks, but their security capabilities in the financial domain have not been explored, and their performance on complex tasks like financial agent remains unknown. This paper presents FinEval, a benchmark designed to evaluate LLMs' financial domain knowledge and practical abilities. The dataset contains 8,351 questions categorized into four different key areas: Financial Academic Knowledge, Financial Industry Knowledge, Financial Security Knowledge, and Financial Agent. Financial Academic Knowledge comprises 4,661 multiple-choice questions spanning 34 subjects such as finance and economics. Financial Industry Knowledge contains 1,434 questions covering practical scenarios like investment research. Financial Security Knowledge assesses models through 1,640 questions on topics like application security and cryptography. Financial Agent evaluates tool usage and complex reasoning with 616 questions. FinEval has multiple evaluation settings, including zero-shot, five-shot with chain-of-thought, and assesses model performance using objective and subjective criteria. Our results show that Claude 3.5-Sonnet achieves the highest weighted average score of 72.9 across all financial domain categories under zero-shot setting. Our work provides a comprehensive benchmark closely aligned with Chinese financial domain.
Cite
@article{arxiv.2308.09975,
title = {FinEval: A Chinese Financial Domain Knowledge Evaluation Benchmark for Large Language Models},
author = {Xin Guo and Haotian Xia and Zhaowei Liu and Hanyang Cao and Zhi Yang and Zhiqiang Liu and Sizhe Wang and Jinyi Niu and Chuqi Wang and Yanhui Wang and Xiaolong Liang and Xiaoming Huang and Bing Zhu and Zhongyu Wei and Yun Chen and Weining Shen and Liwen Zhang},
journal= {arXiv preprint arXiv:2308.09975},
year = {2024}
}