FSE Proceedings of the 34th ACM International Conference on the Foundations of Software Engineering2026匹配度 5/10 · 论文研究的是 LLM 驱动的缺陷修复系统（含自动生成测试、代码-测试联合迭代）在 SWE-bench 上的测试过拟合现象，属于“Agentic 程序”评测与可靠性话题，与用户关注的 Coding Agent 行为分析有交集；但它既不涉及静态分析技术，也不涉及浮点或贝叶斯程序分析，且更偏软件工程实证/基准评测而非 PL 形式化与工具落地。引用 5inbox
Toufique Ahmed, Jatin Ganhotra, Avraham Shinnar, Martin Hirzel
Tests can be useful towards resolving issues on code repositories. However, relying too much on tests for issue resolution can lead to code that technically passes observed tests but actually misses important cases or even breaks functionality. This problem, called test overfitting, is exacerbated by the fact that issues usually lack readily executable tests. Instead, several issue resolution systems use tests auto-generated from issues, which may be imperfect. Some systems even iteratively refine code and tests jointly. This paper presents the first empirical study of test overfitting in this setting.
@inproceedings{AhmedGSH26,
title = {Investigating Test Overfitting on SWE-bench},
author = {Toufique Ahmed and Jatin Ganhotra and Avraham Shinnar and Martin Hirzel},
booktitle = {FSE},
year = {2026},
doi = {10.1145/3803437.3805574}
}