Work

Follow the Path: Reasoning over Knowledge Graph Paths to Improve Large Language Model Factuality

Year: 2026

Type: conference-paper

Abstract: We introduce fs1, a simple yet effective method that improves the factuality of reasoning traces by collecting them from large reasoning models and grounding them in knowledge graph (KG) paths.We fine-tune eight instructiontuned Large Language Models (LLMs) on 3.9K factually grounded reasoning traces and rigorously evaluate them on six complex opendomain question-answering (QA) benchmarks encompassing 23.9K questions.Our results demonstrate that our fs1-tuned model consistently outperforms instruction-tuned counterparts with parallel sampling by 6-14 absolute points (pass@16).Our detailed analysis shows that fs1 considerably improves model performance over more complex questions (requiring 3 or more hops on KG paths) and numerical answer types compared to the baselines.Furthermore, in single-pass inference, we notice that smaller LLMs show the most improvements.While prior works demonstrate the effectiveness of reasoning traces primarily in the STEM domains, our work shows strong evidence that anchoring reasoning to factual KG paths is a critical step in transforming LLMs for reliable knowledge-intensive tasks.

Source: Findings of the Association for Computational Linguistics: ACL (Findings of ACL)

DOI: 10.18653/v1/2026.findings-acl.561

Authors: Mike Zhang, Johannes Bjerva, Russa Biswas

Institutions: University of Copenhagen, Aalborg University

Language: English

Topic: Topic Modeling

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

Funders: European High Performance Computing Joint Undertaking, Villum Fonden, China Scholarship Council, Danish e-Infrastructure Cooperation

Award: G6736582538