Few-shot Reranking for Multi-hop QA via Language Model Prompting

TL;DR

This paper introduces PromptRank, a few-shot reranking method for multi-hop question answering that uses large language model prompting to achieve high performance with minimal training data.

Contribution

PromptRank is a novel approach that leverages language model prompts for multi-hop path reranking, reducing the need for extensive labeled data.

Findings

PromptRank achieves 73.6 recall@10 on HotpotQA with only 128 training examples.

It outperforms existing methods trained on thousands of examples.

PromptRank demonstrates strong few-shot performance in multi-hop QA reranking.

Abstract

We study few-shot reranking for multi-hop QA with open-domain questions. To alleviate the need for a large number of labeled question-document pairs for retriever training, we propose PromptRank, which relies on large language models prompting for multi-hop path reranking. PromptRank first constructs an instruction-based prompt that includes a candidate document path and then computes the relevance score between a given question and the path based on the conditional likelihood of the question given the path prompt according to a language model. PromptRank yields strong retrieval performance on HotpotQA with only 128 training examples compared to state-of-the-art methods trained on thousands of examples -- 73.6 recall@10 by PromptRank vs. 77.8 by PathRetriever and 77.5 by multi-hop dense retrieval. Code available at https://github.com/mukhal/PromptRank