A new research paper investigates whether standard code evaluation metrics can effectively detect plagiarism in source code. The study compared five metrics—CodeBLEU, CrystalBLEU, RUBY, TSED, and CodeBERTScore—against specialized plagiarism detection tools like JPlag and Dolos. Results indicate that these metrics can be comparable to dedicated tools, especially CrystalBLEU when combined with preprocessing, and perform best on less modified code. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT Evaluates the effectiveness of code evaluation metrics for plagiarism detection, potentially impacting academic integrity tools.
RANK_REASON This is a research paper evaluating existing metrics for a specific task.