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

2026-05-02

AI Summary

This work addresses the challenge of effectively integrating retrieved information with the reasoning capabilities of large language models in traditional retrieval-augmented generation (RAG) approaches. The authors propose Verbal-R3, a novel framework that introduces Verbal Annotations—analytical narratives explicitly describing the logical relationship between queries and retrieved passages—as a bridge between retrieval and reasoning. Verbal-R3 features an iterative retrieve-and-reason generator and a verbal reranker that jointly outputs relevance scores and Verbal Annotations. It further incorporates a relevance-guided test-time scaling strategy to dynamically allocate computational resources and optimize reasoning trajectories. Evaluated on multiple complex question-answering benchmarks, Verbal-R3 achieves state-of-the-art performance, significantly improving both answer accuracy and contextual consistency.

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.

Problem

• Retrieval-Augmented Generation
• Large Language Model
• Information Integration
• Reasoning
• Question Answering

Innovation

• Verbal Reranker
• Verbal Annotations
• Retrieval-Augmented Generation
• Test-Time Scaling
• Agentic RAG