This repository curates essential research papers on using Large Language Models (LLMs) as optimization engines and for automating prompt engineering. It targets researchers and practitioners in AI and machine learning, providing a centralized resource to understand the state-of-the-art in LLM-driven optimization techniques. The benefit is a streamlined path to understanding advanced LLM applications in solving complex optimization problems.

How It Works

The project itself is a curated list of academic papers. The papers themselves explore various methodologies for leveraging LLMs in optimization. These include treating LLMs as gradient-free optimizers, employing reinforcement learning for prompt tuning, using evolutionary algorithms for prompt evolution, and developing frameworks for LLMs to generate or optimize solutions for combinatorial problems. The novelty lies in reframing LLMs not just as text generators but as powerful tools capable of complex reasoning and optimization tasks.

Quick Start & Requirements

This repository is a curated list of research papers and does not contain executable code or a direct installation process. Users are expected to access the papers via the provided links (e.g., [abs], [openreview]) to read them.

Highlighted Details

• Covers a broad spectrum of LLM optimization techniques, from prompt engineering automation (RLPrompt, GrIPS) to LLMs acting as optimizers themselves (Large Language Models as Optimizers, Eureka).
• Includes papers exploring LLMs for specific domains like code generation (Self-Taught Optimizer, LLaMoCo) and business optimization (AI-Copilot for Business Optimisation).
• Features recent preprints (up to October 2023 and early 2024), indicating a focus on cutting-edge research.
• References related resources like "awesome-fm4co" for Foundation Models in Combinatorial Optimization.

Maintenance & Community

The repository was created on 2023-10-06. No specific maintainers, community links (like Discord/Slack), or roadmap details are provided in the README.

Licensing & Compatibility

No licensing information is provided for the repository itself or the curated list. Users should refer to the individual licenses of the linked research papers.

Limitations & Caveats

This is a static list of papers and does not provide code implementations, benchmarks, or direct access to the models discussed. The rapid pace of LLM research means the list may become outdated quickly. Users must independently evaluate the methodologies and findings presented in the papers.