Guardado en:
Detalles Bibliográficos
Autores principales: Hu, Junwei, Zheng, Weicheng, Liu, Yihan, Liu, Yan
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2504.15989
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912403206176768
author Hu, Junwei
Zheng, Weicheng
Liu, Yihan
Liu, Yan
author_facet Hu, Junwei
Zheng, Weicheng
Liu, Yihan
Liu, Yan
contents With the increasing adoption of large language models (LLMs) in software engineering, the Chain of Thought (CoT) reasoning paradigm has become an essential approach for automated code repair. However, the explicit multi-step reasoning in CoT leads to substantial increases in token consumption, reducing inference efficiency and raising computational costs, especially for complex code repair tasks. Most prior research has focused on improving the correctness of code repair while largely overlooking the resource efficiency of the reasoning process itself. To address this challenge, this paper proposes three targeted optimization strategies: Context Awareness, Responsibility Tuning, and Cost Sensitive. Context Awareness guides the model to focus on key contextual information, Responsibility Tuning refines the structure of the reasoning process through clearer role and responsibility assignment, and Cost Sensitive incorporates resource-awareness to suppress unnecessary token generation during inference. Experiments across diverse code repair scenarios demonstrate that these methods can significantly reduce token consumption in CoT-based reasoning without compromising repair quality. This work provides novel insights and methodological guidance for enhancing the efficiency of LLM-driven code repair tasks in software engineering.
format Preprint
id arxiv_https___arxiv_org_abs_2504_15989
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Optimizing Token Consumption in LLMs: A Nano Surge Approach for Code Reasoning Efficiency
Hu, Junwei
Zheng, Weicheng
Liu, Yihan
Liu, Yan
Software Engineering
With the increasing adoption of large language models (LLMs) in software engineering, the Chain of Thought (CoT) reasoning paradigm has become an essential approach for automated code repair. However, the explicit multi-step reasoning in CoT leads to substantial increases in token consumption, reducing inference efficiency and raising computational costs, especially for complex code repair tasks. Most prior research has focused on improving the correctness of code repair while largely overlooking the resource efficiency of the reasoning process itself. To address this challenge, this paper proposes three targeted optimization strategies: Context Awareness, Responsibility Tuning, and Cost Sensitive. Context Awareness guides the model to focus on key contextual information, Responsibility Tuning refines the structure of the reasoning process through clearer role and responsibility assignment, and Cost Sensitive incorporates resource-awareness to suppress unnecessary token generation during inference. Experiments across diverse code repair scenarios demonstrate that these methods can significantly reduce token consumption in CoT-based reasoning without compromising repair quality. This work provides novel insights and methodological guidance for enhancing the efficiency of LLM-driven code repair tasks in software engineering.
title Optimizing Token Consumption in LLMs: A Nano Surge Approach for Code Reasoning Efficiency
topic Software Engineering
url https://arxiv.org/abs/2504.15989