Is there a way to use large language model to realize query focused multi-document summarization?

Hi, I am a beginner in nlp and I just learnt attention mechanism. I am wondering if there is a way to use established language models to apply the idea of attention to query-based multi-document search and summarization? Specifically, I want to first search a query and get a list of text segments which have the top n similarity scores with the query. Then I want to summarize information from these search results by evaluating their similarity scores and extracting more information from the segments with higher similarity scores. Is there a established model or framework could realize this? Or any papers or projects for reference? Thanks a lot.