This repository curates essential academic papers, insightful blogs, and API tools related to the pre-training and fine-tuning methodologies behind ChatGPT. It serves as a valuable resource for researchers, developers, and practitioners seeking to understand and replicate advanced language model capabilities.

How It Works

The project acts as a structured bibliography, categorizing key research papers that define the evolution of large language models, from GPT-1 to GPT-4, and their associated training techniques like Reinforcement Learning from Human Feedback (RLHF) and Proximal Policy Optimization (PPO). It also links to relevant blogs and API implementations, providing a comprehensive overview of the ChatGPT ecosystem.

Quick Start & Requirements

• No installation or execution commands are provided, as this is a curated list of resources.
• Requirements are implicit: access to academic papers (often via arXiv or publisher sites) and the ability to explore linked blogs and API projects.
• Links to official ChatGPT blogs and prompting guides are included.

Highlighted Details

• Comprehensive list of foundational papers from GPT-1 to GPT-4.
• Inclusion of key papers on fine-tuning and alignment techniques (InstructGPT, RLHF, RAFT).
• Links to related LLM projects like LMFlow, DetGPT, and Toolformer.
• Curated list of unofficial ChatGPT API wrappers across Python, JavaScript/TypeScript, and Dart.

Maintenance & Community

• The repository is maintained by shizhediao.
• Links to official ChatGPT blogs and prompting guides are provided.

Licensing & Compatibility

• The repository itself does not appear to have a specific license declared in the README.
• The linked papers and API projects will have their own respective licenses.

Limitations & Caveats

This repository is a curated list and does not provide code for training or running models. The API links are explicitly marked as "Non-Official," indicating potential instability or changes.