AlbertMaxwell/MMSNews-QA

MMSNews-QA数据集经过精心策划，旨在评估大型语言模型（LLMs）在聚合和分析加密货币新闻方面的能力，支持在动态数字资产领域的全面研究和战略决策。数据集包括新闻文章的日期、时间戳、全文、类别、关联的加密货币、交易符号、不同场景的回报值、真实标签和最大回报值等特征。数据集分为训练集和测试集，分别包含1144和128个示例，总最大回报值分别为1745.76%和530.34%。

The MMSNews-QA dataset is meticulously curated to assess the ability of LLMs to aggregate and analyze cryptocurrency news, empowering comprehensive research and strategic decision-making in the dynamic digital asset landscape. The dataset includes features such as date, timestamp, news text, category, associated cryptocurrency, trading symbol, return values for different scenarios, ground truth label based on maximum return, and maximum return value. The dataset is divided into a training set with 1144 examples and a testing set with 128 examples, with total maximum returns of 1745.76% and 530.34% respectively. The dataset also includes data visualizations such as cumulative performance, distribution of max return, sum of max return per category, and ground truth distribution per category. The dataset is proprietary and not for public use.