Rethinking All Evidence: Enhancing Trustworthy Retrieval-Augmented Generation via Conflict-Driven Summarization

TL;DR

This paper introduces CARE-RAG, a framework that enhances retrieval-augmented generation by conflict-driven summarization of evidence, improving trustworthiness and reliability in the presence of conflicting or noisy information.

Contribution

It proposes a novel conflict-aware summarization method that synthesizes internal and retrieved evidence, significantly improving RAG performance on noisy or conflicting data.

Findings

Outperforms strong RAG baselines on revised QA datasets.

Effectively detects and summarizes conflicts in evidence.

Improves trustworthiness in retrieval-augmented generation.

Abstract

Retrieval-Augmented Generation (RAG) enhances large language models (LLMs) by integrating their parametric knowledge with external retrieved content. However, knowledge conflicts caused by internal inconsistencies or noisy retrieved content can severely undermine the generation reliability of RAG systems.In this work, we argue that LLMs should rethink all evidence, including both retrieved content and internal knowledge, before generating responses.We propose CARE-RAG (Conflict-Aware and Reliable Evidence for RAG), a novel framework that improves trustworthiness through Conflict-Driven Summarization of all available evidence.CARE-RAG first derives parameter-aware evidence by comparing parameter records to identify diverse internal perspectives. It then refines retrieved evidences to produce context-aware evidence, removing irrelevant or misleading content. To detect and summarize conflicts, we distill a 3B LLaMA3.2 model to perform conflict-driven summarization, enabling reliable synthesis across multiple sources.To further ensure evaluation integrity, we introduce a QA Repair step to correct outdated or ambiguous benchmark answers.Experiments on revised QA datasets with retrieval data show that CARE-RAG consistently outperforms strong RAG baselines, especially in scenarios with noisy or conflicting evidence.