Investigating Test Overfitting on SWE-bench
This addresses a critical reliability issue for developers and automated code repair systems, though it is incremental as it focuses on empirical analysis rather than a new solution.
The paper investigates test overfitting in code issue resolution, where systems rely on auto-generated tests that may be imperfect, leading to code that passes tests but misses important cases or breaks functionality. It presents the first empirical study of this problem in the context of SWE-bench.
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.