.png){kind=icon}
.png){kind=icon}
Return
A comparative analysis of the role of Large Language Models and Low-Rank Adaptation in Automated Program Repair
.png){kind=icon}
OA
AI
C
P
A
M
DOI：10.1016/j.infsof.2026.108238.png)
.png){kind=icon}
.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Unified framework to compare linear vs. iterative LLM-based APR (base and LoRA) on Defects4J and HumanEval-Java. • Iterative pipelines markedly increase plausible and correct patches while keeping token and time costs controlled. • LoRA improves semantic correctness as a parameter-efficient accelerator with minimal resource overhead. • Open-source LLMs plus LoRA and iterative feedback achieve competitive repair performance with shorter, cheaper patches.
Keywords:
Automated Program Repair
Large Language Models
Parameter-efficient fine-tuning
LoRA
Iterative repair pipelines
AI Summary
Key information extracted from the uploaded paper, including a brief overview, abstract, background, key highlights, visual analysis, and future outlook.
Journal
IF:
4.3
Papers:
3.7K
Citations:
7.7K
Organization
No organization information available
Cited Papers