ClueAnchor: Clue-Anchored Knowledge Reasoning Exploration and Optimization for Retrieval-Augmented Generation

TL;DR

ClueAnchor enhances retrieval-augmented generation by extracting key clues from retrieved documents, exploring multiple reasoning paths, and optimizing for more faithful, interpretable, and robust knowledge-based responses.

Contribution

It introduces a novel clue-anchored reasoning framework that improves the utilization of retrieved knowledge in LLMs for better factuality and interpretability.

Findings

Significantly outperforms prior RAG baselines in reasoning completeness and robustness.

Demonstrates resilience to noisy or partially relevant retrieved content.

Capable of identifying supporting evidence without explicit clue supervision.

Abstract

Retrieval-Augmented Generation (RAG) augments Large Language Models (LLMs) with external knowledge to improve factuality. However, existing RAG systems frequently underutilize the retrieved documents, failing to extract and integrate the key clues needed to support faithful and interpretable reasoning, especially in cases where relevant evidence is implicit, scattered, or obscured by noise. To address this issue, we propose ClueAnchor, a novel framework for enhancing RAG via clue-anchored reasoning exploration and optimization. ClueAnchor extracts key clues from retrieved content and generates multiple reasoning paths based on different knowledge configurations, optimizing the model by selecting the most appropriate reasoning path for the given context through reward-based preference optimization. Experiments show that ClueAnchor significantly outperforms prior RAG baselines in the completeness and robustness of reasoning. Further analysis confirms its strong resilience to noisy or partially relevant retrieved content, as well as its capability to identify supporting evidence even in the absence of explicit clue supervision during inference. All codes are available at https://github.com/thunlp/ClueAnchor.