BriefGPT.xyz
Oct, 2023
搜索适配器:信息检索的文本嵌入定制化
Search-Adaptor: Text Embedding Customization for Information Retrieval
HTML
PDF
Jinsung Yoon, Sercan O Arik, Yanfei Chen, Tomas Pfister
TL;DR
通过自定义 Search-Adaptor 方法,我们可以提高预训练大型语言模型在信息检索和搜索领域的性能,有效且稳健地修改原始文本嵌入向量,与任何语言模型集成,并在多个实际英语和多语种检索数据集上取得一致且显著的性能提升。
Abstract
text embeddings
extracted by pre-trained
large language models
(LLMs) have significant potential to improve
information retrieval
and
→