A RAG Method for Source Code Inquiry Tailored to Long-Context LLMs

Abstract

Although the context length limitation of large language models (LLMs) has been mitigated, it still hinders their application to software development tasks. This study proposes a method incorporating execution traces into RAG for inquiries about source code. Small-scale experiments confirm a tendency for the method to contribute to improving LLM response quality.

6 pages, 2 columns, English translation of the manuscript originally presented in Japanese at a domestic workshop