Test-Time Strategies for More Efficient and Accurate Agentic RAG

TL;DR

This paper proposes test-time modifications to Search-R1 in Retrieval-Augmented Generation systems, introducing contextualization and de-duplication modules to enhance accuracy and efficiency in complex question answering tasks.

Contribution

It introduces novel test-time strategies, including a contextualization module and a de-duplication approach, to improve reasoning and reduce retrieval redundancy in agentic RAG systems.

Findings

Achieved a 5.6% increase in EM score on HotpotQA and Natural Questions datasets.

Reduced the number of retrieval turns by 10.5%, improving efficiency.

Enhanced answer correctness through the integration of GPT-4.1-mini for contextualization.

Abstract

Retrieval-Augmented Generation (RAG) systems face challenges with complex, multihop questions, and agentic frameworks such as Search-R1 (Jin et al., 2025), which operates iteratively, have been proposed to address these complexities. However, such approaches can introduce inefficiencies, including repetitive retrieval of previously processed information and challenges in contextualizing retrieved results effectively within the current generation prompt. Such issues can lead to unnecessary retrieval turns, suboptimal reasoning, inaccurate answers, and increased token consumption. In this paper, we investigate test-time modifications to the Search-R1 pipeline to mitigate these identified shortcomings. Specifically, we explore the integration of two components and their combination: a contextualization module to better integrate relevant information from retrieved documents into reasoning, and a de-duplication module that replaces previously retrieved documents with the next most relevant ones. We evaluate our approaches using the HotpotQA (Yang et al., 2018) and the Natural Questions (Kwiatkowski et al., 2019) datasets, reporting the exact match (EM) score, an LLM-as-a-Judge assessment of answer correctness, and the average number of turns. Our best-performing variant, utilizing GPT-4.1-mini for contextualization, achieves a 5.6% increase in EM score and reduces the number of turns by 10.5% compared to the Search-R1 baseline, demonstrating improved answer accuracy and retrieval efficiency.