.png){kind=icon}
.png){kind=icon}
Return
Can federated fine-tuning improve LLM-based code summarization? An empirical investigation on open-source repositories
.png){kind=icon}
PRE
AI
J
M
L
S
DOI：10.1016/j.jss.2026.113010.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Code summarization is essential for helping developers navigate complex projects. Recent advancements utilizing Large Language Models (LLMs) reveal limitations in the applicability of vanilla LLMs for proprietary code, motivating further fine-tuning on private datasets. Additionally, the siloed nature of code within institutions makes centralized fine-tuning impractical.
Journal
IF:
4.1
Papers:
5.4K
Citations:
8.4K
Organization
No organization information available
Cited Papers