An empirical study of four open-source LLMs across six benchmarks shows code-specialized models often outperform larger general models, and most correct repairs appear early in generation.
S3: syntax- and semantic-guided repair synthesis via programming by examples,
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
-
Empirical Evaluation of Large Language Models in Automated Program Repair
An empirical study of four open-source LLMs across six benchmarks shows code-specialized models often outperform larger general models, and most correct repairs appear early in generation.