This repository contains individual research projects, each representing an exploration driven by an LLM tool, predominantly Claude Code. Each project's entire codebase and accompanying documentation were generated by an LLM, demonstrating its capabilities in various research contexts. The primary goal is to document and share the process of using LLMs to conduct research, highlighting the steps from prompt engineering to code and documentation generation.
The project is unique in its comprehensive documentation of research conducted entirely by an LLM. It offers insights into prompt design and the resulting code and analysis. The repository's well-structured organization, with each project residing in a separate directory, facilitates easy exploration and understanding. The inclusion of links to relevant prompts and transcripts enables reproducibility and detailed examination of the research process.
- Code Generation: Demonstrates the LLM's ability to generate functional code for research purposes.
- Documentation: Includes generated documentation explaining the research process and findings.
- Prompt Engineering: Provides examples of prompts used to guide the LLM's research activities.
- Reproducibility: Aims to facilitate the reproduction of research based on the provided code and prompts.
- LLM Tooling: Showcases the utilization of LLMs as research assistants for code analysis and documentation.
The project is actively maintained with new research projects added regularly. The README provides clear instructions and links to relevant resources. Ongoing efforts are dedicated to incorporating AI-generated summaries for each project. The repository benefits from a consistent structure and a clear commitment to documenting the entire AI-driven research workflow.
This repository is valuable for researchers, developers, and anyone interested in exploring the potential of LLMs in research. It provides practical examples of how LLMs can be used to generate code, documentation, and analysis. It's a resource for understanding how to leverage these tools for rapid prototyping and knowledge discovery, offering an alternative to more traditional research methodologies.