Better Patching Using LLM Prompting, via Self-Consistency

Large Language models (LLMs) can be induced to solve non-trivial problems with “few-shot” prompts including illustrative problem-solution examples. Now if the few-shots also include “chain of thought” ( <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathcal{C}oT$</tex> ) explanations, which are of the form problem-explanation-solution, LLMs will generate a “explained” solution, and perform even better. Recently an exciting, substantially better technique, self-consistency [1] ( <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathcal{S}-C$</tex> ) has emerged, based on the intuition that there are many plausible explanations for the right solution; when the LLM is sampled repeatedly to generate a pool of explanation-solution pairs, for a given problem, the most frequently occurring solutions in the pool (ignoring the explanations) tend to be even more likely to be correct! Unfortunately, the use of this highly-performant <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathcal{S}-C$</tex> (or even <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathcal{C}oT$</tex> ) approach in software engineering settings is hampered by the lack of explanations; most software datasets lack explanations. In this paper, we describe an application of the <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$\mathcal{S}-C$</tex> approach to program repair, using the commit log on the fix as the explanation, only in the illustrative few-shots. We achieve state-of-the art results, beating previous approaches to prompting-based program repair, on the MODIT dataset; we also find evidence suggesting that the correct commit messages are helping the LLM learn to produce better patches.