Word-level adversarial changes in prompts, code, and comments degrade code-generation correctness more than sentence-level rewrites, but classification errors and missing error bars weaken the claim.
Title resolution pending
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
-
Adversarial Attack Classification and Robustness Testing for Large Language Models for Code
Word-level adversarial changes in prompts, code, and comments degrade code-generation correctness more than sentence-level rewrites, but classification errors and missing error bars weaken the claim.