ModelBloombergGPT
Bloomberg GPT
BloombergGPT是由彭博社开发的金融领域专用大型语言模型，专为金融行业自然语言处理任务设计。该模型基于金融新闻、证券文件、研究报告等大规模金融专有语料从头训练，能够处理情感分析、命名实体识别、金融问答等专业任务，在金融文本理解方面具有较强的领域针对性，是通用大语言模型在垂直金融场景下的专业化实践代表。
All Facts (2)
Unverified
Bloomberg GPT's training data includes 40 years of financial news, SEC filings, research reports, and more, with the project consuming tens of millions of dollars in compute costs
60%UnverifiedBloomberg GPT is a large language model with 50 billion parameters released in March 2023, trained from scratch by the Bloomberg team using proprietary financial domain data
60%