About this document

Comment Traps: How Defective Commented-Out Code Augment Defects in AI-Assisted Code Generation by zhufengning18 is a document available to read on EtoBox.

This study investigates the impact of defective commented-out code (CO code) on the performance of AI coding assistants, specifically GitHub Copilot and Cursor, revealing that such defects can lead to a significant increase in generated code defects. Experimental results indicate that up to 58.17% of generated code contains defects when influenced by defective CO code, and attempts to mitigate this through prompt engineering yield limited success. The findings highlight the need for improved robustness and

Author

zhufengning18

Language

EN