Foundations of LLMs is an open-source educational textbook that introduces the fundamental principles, architectures, and techniques behind large language models. Developed by researchers at Zhejiang University, it provides structured explanations of both established methods and emerging AI technologies. The material covers statistical language models, recurrent neural networks, Transformers, and modern language model architectures. Dedicated chapters explore prompt engineering, in-context learning, and chain-of-thought reasoning. Readers also learn parameter-efficient fine-tuning, model editing, and retrieval-augmented generation. The repository includes downloadable textbook chapters, an English edition, diagrams, and collections of relevant academic papers. It serves students, researchers, and developers seeking a systematic understanding of language model development.
Features
- Comprehensive Large Language Model Textbook
- Transformer and Neural Network Architecture Coverage
- Prompt Engineering and Reasoning Techniques
- Parameter-Efficient Fine-Tuning Methods
- Model Editing and Retrieval-Augmented Generation
- Model Editing and Retrieval-Augmented Generation