Probing-RAG: Self-Probing to Guide Language Models in Selective Document Retrieval

TL;DR

Probing-RAG introduces a method that uses internal model states to decide when additional document retrieval is needed, improving efficiency and accuracy in open-domain question answering.

Contribution

It presents a novel approach leveraging hidden states for adaptive retrieval decisions, enhancing retrieval efficiency in RAG systems.

Findings

Outperforms previous methods on five QA datasets.

Reduces redundant retrieval steps.

Improves retrieval decision reliability.

Abstract

Retrieval-Augmented Generation (RAG) enhances language models by retrieving and incorporating relevant external knowledge. However, traditional retrieve-and-generate processes may not be optimized for real-world scenarios, where queries might require multiple retrieval steps or none at all. In this paper, we propose a Probing-RAG, which utilizes the hidden state representations from the intermediate layers of language models to adaptively determine the necessity of additional retrievals for a given query. By employing a pre-trained prober, Probing-RAG effectively captures the model's internal cognition, enabling reliable decision-making about retrieving external documents. Experimental results across five open-domain QA datasets demonstrate that Probing-RAG outperforms previous methods while reducing the number of redundant retrieval steps.