Why Neighborhoods Matter: Traversal Context and Provenance in Agentic GraphRAG

TL;DR

This paper investigates how external evidence and graph traversal context influence answer accuracy and citation faithfulness in Agentic GraphRAG systems, emphasizing the importance of provenance beyond source support.

Contribution

It introduces a trajectory-level perspective on citation faithfulness, highlighting the roles of visited but uncited entities and graph structure in answer generation.

Findings

Cited evidence is often necessary for accurate answers.

Uncited traversal context and graph structure also impact answer correctness.

Citation evaluation should consider the entire retrieval trajectory, not just source support.

Abstract

Retrieval-Augmented Generation can improve factuality by grounding answers in external evidence, but Agentic GraphRAG complicates what it means for citations to be faithful. In these systems, an agent explores a knowledge graph before producing an answer and a small set of citations. We frame citation faithfulness as a trajectory-level problem: final citations should not only support the answer, but also account for the graph traversal, structure, and visited-but-uncited entities that may influence it. Through controlled ablation experiments, we compare the effects of isolating, removing, and masking cited and uncited graph entities. Our results show that cited evidence is often necessary, as removing it substantially changes answers and reduces accuracy. However, citations are not sufficient, because accurate answers can also depend on uncited traversal context and surrounding graph structure. These findings suggest that citation evaluation in Agentic GraphRAG should move beyond source support toward provenance over the broader retrieval trajectory.