LLM4SZZ enhances the industry-standard SZZ algorithm for identifying bug-inducing commits by leveraging large language models to provide context-aware code analysis.

• Improves bug detection accuracy by analyzing semantic relationships rather than relying solely on static techniques
• Incorporates context-enhanced assessment to better understand code changes and their relationship to bugs
• Demonstrates significant performance improvement over traditional SZZ variants
• Provides a more reliable foundation for bug prediction and static code analysis tools

This research matters for engineering teams by offering more accurate identification of bug sources, enabling better resource allocation for code review and maintenance, and improving overall software quality assurance processes.

LLM4SZZ: Enhancing SZZ Algorithm with Context-Enhanced Assessment on Large Language Models