Year: 2026
Type: conference-paper
Abstract: Large language models (LLMs) have shown strong performance in zero-shot summarization, but often struggle to model document structure and identify salient information in long texts.In this work, we introduce Struc-Sum, a training-free prompting framework that enhances LLM reasoning through sentencelevel graph structures.StrucSum injects structural signals into prompts via three targeted strategies: Neighbor-Aware Prompting (NAP) for local context, Centrality-Aware Prompting (CAP) for importance estimation, and Centrality-Guided Masking (CGM) for efficient input reduction.Experiments on ArXiv, PubMed, and Multi-News demonstrate that StrucSum consistently improves both summary quality and factual consistency over unsupervised baselines and vanilla prompting.In particular, on ArXiv, it increases FactCC and Sum-maC by 19.2% and 8.0% points, demonstrating stronger alignment between summaries and source content.The ablation study shows that the combination of multiple strategies does not yield clear performance gains; therefore, structure-aware prompting with graph-based information represents a promising and underexplored direction for the advancement of zeroshot extractive summarization with LLMs.Our source code is publicly available.
Authors: Haohan Yuan, Sukhwa Hong, Haopeng Zhang
Institutions: University of Hawaiʻi at Mānoa, University of Hawaii at Hilo
Language: English
Topic: Topic Modeling
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
Funder: National Science Foundation
OpenAlex ID: W7140131920