Search-Adaptor: Embedding Customization for Information Retrieval

arXiv:2310.08750v2 Announce Type: replace
Abstract: Embeddings extracted by pre-trained Large Language Models (LLMs) have significant potential to improve information retrieval and search. Beyond the zero-shot setup in which they are being conventionally used, being able to take advantage of the information from the relevant query-corpus paired data can further boost the LLM capabilities. In this paper, we propose a novel method, Search-Adaptor, for customizing LLMs for information retrieval in an efficient and robust way. Search-Adaptor modifies the embeddings generated by …

abstract arxiv beyond boost capabilities cs.lg customization data embedding embeddings information language language models large language large language models llm llms query retrieval search setup the information type zero-shot