About AI Notes

AI Notes is a public, source-linked collection of explanations about language models, AI agents, retrieval-augmented generation, and evaluation. It is maintained by Donghyun Lee as an educational resource and an evolving record of technical study.

The site favors compact explanations that connect concepts to practical system design. It is not intended to replace primary papers, official documentation, or model-specific safety guidance. References are provided so readers can check important claims at their source.

Editorial Principles

• Explain the mechanism, not only the terminology.
• Separate model behavior from application and infrastructure behavior.
• State important limitations and failure modes.
• Prefer public primary sources and official documentation.
• Treat evaluation and evidence as part of system design.
• Revise pages when the underlying tools or evidence change.

Scope

The current scope covers LLM fundamentals, agent and RAG patterns, common implementation tools, and evaluation. Personal notes, private project records, raw source archives, and operational material are outside the scope of this site.