Large Language Model Critics for Execution-Free Evaluation of Code Changes
This work addresses the need for better evaluation metrics in automated software engineering, offering a tool for assessing code changes without execution, which is incremental but improves upon existing sparse metrics.
The paper tackles the problem of evaluating code changes in LLM-based software engineering workflows by designing LLM-based critics that provide execution-free evaluation proxies, achieving an F1 score of 91.6% for predicting executability and 84.8% accuracy for build status prediction in SWE-bench.
Large language models (LLMs) offer a promising way forward for automating software engineering tasks, such as bug fixes, feature additions, etc., via multi-step LLM-based agentic workflows. However, existing metrics for evaluating such workflows, mainly build status and occasionally log analysis, are too sparse and limited in providing the information needed to assess the quality of changes made. In this work, we designed LLM-based critics to derive well-structured and rigorous intermediate/step-level, execution-free evaluation proxies for repo-level code changes. Importantly, we assume access to the gold test patch for the problem (i.e., reference-aware) to assess both semantics and executability of generated patches. With the gold test patch as a reference, we predict executability of all editing locations with an F1 score of 91.6%, aggregating which, we can predict the build status in 84.8% of the instances in SWE-bench. In particular, such an execution-focused LLM critic outperforms other reference-free and reference-aware LLM critics by 38.9% to 72.5%. Moreover, we demonstrate the usefulness of such a reference-aware framework in comparing patches generated by different agentic workflows. Finally, we open-source the library developed for this project, which allows further usage for either other agentic workflows or other benchmarks. The source code is available at https://github.com/amazon-science/code-agent-eval.