Chain of Understanding: Supporting Code Understanding with Large Language Models

2025-04-06

AI Summary

Code auditing poses significant comprehension barriers for non-expert developers due to the cognitive overhead of navigating complex codebases and formulating effective prompts for large language models (LLMs). Method: This paper proposes a hierarchical, progressive code understanding paradigm enabled by an interactive LLM-based analysis system. The system integrates three core components: (1) CodeMap—a structural code visualization module; (2) context-aware conversational interfaces; and (3) a stepwise guidance engine. Crucially, it introduces the “Chain-of-Understanding” cognitive framework, the first to unify hierarchical reasoning with tightly coupled visualization–dialogue interaction. Contribution/Results: The design substantially reduces prompt engineering effort: user studies show a 62% reduction in manual prompt authoring time compared to both pure-LLM baselines and static visualization tools. It improves auditing efficiency and user engagement, earning consistent endorsement from both expert and novice developers.

Technology Category

• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Code auditing demands a robust understanding of codebases - an especially challenging task for end-user developers with limited expertise. To address this, we conducted formative interviews with experienced auditors and identified a Chain-of-Understanding approach, in which Large Language Models (LLMs) guide developers through hierarchical code comprehension - from high-level overviews to specific functions and variables. Building on this, we incorporated the Chain-of-Understanding concept into CodeMap, a system offering interactive visualizations, stepwise guided analysis, and context-aware chatbot support. Through within-subject user studies with 10 participants of diverse backgrounds and 5 expert and 2 novice interviews, CodeMap proved effective in reducing the manual effort of prompt engineering while enhancing engagement with visualization, outperforming both standalone LLMs and traditional static visualization tools.

Problem

• Supporting code understanding for end-user developers with limited expertise
• Guiding hierarchical code comprehension using Large Language Models (LLMs)
• Reducing manual effort in prompt engineering and enhancing visualization engagement

Innovation

• Hierarchical code comprehension with LLMs
• Interactive visualizations in CodeMap system
• Context-aware chatbot support for guidance