This repository is a curated collection of academic publications and methodologies focused on Large Language Models (LLMs) for Software Engineering (LLM4SE). It aims to provide a comprehensive overview of LLM applications in SE, covering pre-training tasks, downstream tasks, and specific use cases like code generation, testing, and maintenance. The target audience includes researchers and practitioners in both LLMs and software engineering.

How It Works

The repository categorizes LLM research within software engineering across various domains, including requirements engineering, development, testing, maintenance, and management. It lists numerous papers, models (encoder-only, encoder-decoder, decoder-only), and benchmarks, providing a structured landscape of the LLM4SE field. The organization facilitates understanding the breadth and depth of LLM applications and research trends.

Quick Start & Requirements

This repository is a survey and does not have direct installation or execution commands. It serves as a reference guide to existing research.

Highlighted Details

• Comprehensive classification of LLMs in SE, covering over 20 sub-domains.
• Extensive list of recent (2020-2024) academic publications and their associated GitHub repositories.
• Detailed categorization of LLM architectures (encoder-only, encoder-decoder, decoder-only) applied to code.
• Inclusion of key benchmarks and datasets used for evaluating LLMs in SE tasks.

Maintenance & Community

The repository is maintained by iSEngLab and welcomes contributions via GitHub issues and pull requests. It cites a survey paper as its primary reference.

Licensing & Compatibility

The repository itself does not specify a license, but it references academic papers, each with its own licensing and usage terms.

Limitations & Caveats

As a survey, this repository does not provide functional code or tools. Its value is in its comprehensive listing and categorization of existing research, requiring users to consult the original papers for implementation details and specific capabilities.