EMNLP Findings2025
Advancing Reasoning with Off-the-Shelf LLMs: A Semantic Structure Perspective
Pengfei He, Zitao Li, Yue Xing, Yaliang Li, Jiliang Tang, Bolin Ding
摘要
Large Language Models (LLMs) have shown strong capabilities in zero-shot reasoning and generalization to new tasks. However, the zero-shot performance of general LLMs on complex tasks, such as multi-hop reasoning, remains suboptimal, while reasoning LLMs suffer from hallucinations and unfaithfulness. In this paper, to handle these limitations, we introduce a novel structure-oriented analysis method that helps LLMs better understand the question structure and guide the problemsolving process. We demonstrate that existing reasoning strategies, such as Chain-of-Thought and ReAct, significantly benefit from the LLM's inherent understanding of semantic structure. We further ground our method in the theory of probabilistic graphical models to support its effectiveness. To enhance the reasoning process, we augment the structureoriented analysis with refinement and retrieval capabilities, forming a multi-agent reasoning system called Structure-oriented Autonomous Reasoning Agents (SARA). Extensive experiments show that SARA significantly improves zero-shot performance on knowledge-intensive and mathematical tasks. Remarkably, our approach makes a general LLM competitive with dedicated reasoning models in several benchmarks and demonstrates strong robustness against corrupted reasoning paths.
问 Lune
几分钟读透这篇 Paper,不用再花一下午。
Lune 能读完全文,也能读引用它的顶会 Paper,讲清它的结论、证据和局限,每一点都注明出自哪篇 Paper。不用硬啃,也不靠猜。
脉络
它引用的 Paper23
- Large Language Models are Zero-Shot ReasonersTakeshi Kojima, Shixiang Shane Gu, Machel Reid, Yutaka Matsuo 等NeurIPS 2022 · 8,168 次引用
- Measuring Massive Multitask Language UnderstandingDan Hendrycks, Collin Burns, Steven Basart, Andy Zou 等ICLR 2021 · 7,905 次引用
- Reflexion: language agents with verbal reinforcement learningNoah Shinn, Federico Cassano, Ashwin Gopinath, Karthik Narasimhan 等NeurIPS 2023 · 5,828 次引用
- Tree of Thoughts: Deliberate Problem Solving with Large Language ModelsShunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran 等NeurIPS 2023 · 5,068 次引用
- An Explanation of In-context Learning as Implicit Bayesian InferenceSang Michael Xie, Aditi Raghunathan, Percy Liang, Tengyu MaICLR 2022 · 1,030 次引用
相关 Paper
- Structure Guided Prompt: Instructing Large Language Model in Multi-Step Reasoning by Exploring Graph Structure of the TextKewei Cheng, Nesreen K. Ahmed, Theodore L. Willke, Yizhou SunEMNLP 2024 · 6 次引用
- Agent Instructs Large Language Models to be General Zero-Shot ReasonersNicholas Crispino, Kyle Montgomery, Fankun Zeng, Dawn Song 等ICML 2024 · 41 次引用
- Large Language and Reasoning Models are Shallow Disjunctive ReasonersIrtaza Khalid, Amir Masoud Nourollah, Steven SchockaertACL 2025
- DS-MHP: Improving Chain-of-Thought through Dynamic Subgraph-Guided Multi-Hop PathYongqiang Liu, Qiyao Peng, Binrong Liu, Hongtao Liu 等EMNLP Findings 2025 · 2 次引用
- SAR: A Structure-Aligned Reasoning Framework for Temporal Knowledge Graph Question AnsweringQianyi Hu, Jiaxue Liu, Xinhui Tu, Shoujin WangAAAI 2026
