Preprint
Aug 2026
Revisiting Feedback-Driven LLM Code Repair: A Replication and Exploratory Java Extension
The results show that previous conclusions from Python may be sensitive to benchmark construction, feedback representation, and tooling ecosystem, motivating more controlled multilingual benchmarks, and confirm key trends under a partially controlled replication of LLMbased repair systems.
L. Lalonde, Wassim Keddache, Thomas Perron Touchette et al.
· 0 citations