Abstract

Just as instructions have enabled complex user tasks and zero-shot optimization via prompts, so can instructions in retrieval enable complex information needs with zero-shot adaptation. In this talk, we describe efforts to build a promptable retriever that takes in this information like an LM.  We build an evaluation benchmark to test retrieval models and propose a new retriever, Promptriever, which shows large gains on instruction following and is the first retrieval model to be promptable.