Japanese Financial Benchmark Dataset
收藏资源简介:
本研究构建的‘日本金融基准数据集’专为大型语言模型设计,包含五个特定于日本金融领域的基准任务。数据集通过网络爬取和清洗公开文档构建,涵盖情感分析、证券分析基础知识、审计、财务规划师考试及证券经纪人代表测试等多个方面。该数据集旨在评估和提升大型语言模型在处理金融领域日语文本的能力,特别是在理解和生成金融相关内容方面的性能。
The 'Japanese Financial Benchmark Dataset' developed in this study is specifically designed for large language models (LLMs), and includes five benchmark tasks tailored to the Japanese financial domain. The dataset is constructed via web crawling and cleaning of publicly available documents, covering multiple aspects such as sentiment analysis, basic knowledge of securities analysis, auditing, financial planner qualification examinations, and securities broker representative qualification tests. This dataset aims to evaluate and enhance the capabilities of large language models in processing Japanese financial text, particularly their performance in understanding and generating finance-related content.

- 1Construction of a Japanese Financial Benchmark for Large Language ModelsPreferred Networks, Inc. · 2024年



