A systematic review of 45 papers shows metamorphic testing of code models relies mostly on identifier renaming and dead code insertion, targets encoder-only models like CodeBERT, and under-covers generative tasks, newer LLMs, JavaScript and C#, and standardized metrics.
Code Di"erence Guided Adversarial Example Genera- tion for Deep Code Models
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Metamorphic Testing of Deep Code Models: A Systematic Literature Review
A systematic review of 45 papers shows metamorphic testing of code models relies mostly on identifier renaming and dead code insertion, targets encoder-only models like CodeBERT, and under-covers generative tasks, newer LLMs, JavaScript and C#, and standardized metrics.