Verbal-R3: Verbal Reranker as the Missing Bridge between Retrieval and Reasoning

TL;DR

Verbal-R3 introduces a novel framework that uses verbal annotations to explicitly connect retrieval results with reasoning, significantly improving large language model performance on complex question answering tasks.

Contribution

The paper proposes Verbal-R3, a new agentic RAG framework with a generator and verbal reranker that enhances reasoning through explicit verbal annotations and relevance-guided inference.

Findings

Verbal-R3 achieves state-of-the-art results on complex QA benchmarks.

Verbal Annotations substantially improve LLM's ability to generate accurate responses.

Relevance-guided test-time scaling enhances inference efficiency.

Abstract

The conventional Retrieval-Augmented Generation (RAG) paradigm of injecting raw retrieved texts into the Large Language Model (LLM)'s context often results in suboptimal integration of retrieved information. This paper proposes to bridge retrieval results and the LLM's reasoning ability through Verbal Annotations, analytic narratives that explicitly articulate the logical connection between a search query and retrieved contexts. Our empirical investigation reveals the potential of Verbal Annotations to substantially enhance the LLM's ability to generate accurate, contextually-grounded responses. Motivated by this finding, we introduce Verbal-R3, a novel agentic RAG framework that consists of a Generator and a Verbal Reranker. The Generator performs iterative retrieval and reasoning, while the Verbal Reranker returns relevance scores and Verbal Annotations to guide the reasoning and answering process of the Generator. The inference process of Verbal-R3 is further refined through relevance-guided test-time scaling, which efficiently allocates test-time compute for effective trajectory expansion. Verbal-R3 achieves state-of-the-art performance on complex Question Answering benchmarks, validating the effectiveness of the proposed framework.