Why this matters now
Large language models have moved from isolated research artifacts to the core toolkit for many NLP systems — yet most learners face a fragmented landscape of papers, blog posts, and tooling. This repository offers a single, curated textbook-style path that connects classical language-model theory to today’s LLM engineering practices, making it easier to see the "so what" behind each technique.
What Sets It Apart
- Chapter-driven, syllabus-ready structure: the material is organized into focused chapters (language-model basics, LLM architecture evolution, prompt engineering, parameter-efficient fine-tuning, model editing, and retrieval-augmented generation), so instructors can adopt individual chapters or the full sequence.
- Curated paper lists per chapter: each section includes a maintained list of influential and recent papers, reducing the discovery cost for students and researchers and helping track fast-moving progress without reading everything.
- Bilingual and modular assets with monthly updates: the repo provides Chinese and English chapter PDFs and per-chapter files that are updated regularly, making it practical for bilingual courses and incremental syllabus improvements.
- Community-maintained by an academic lab: maintained by the DAILY Lab (Zhejiang University), with contact info in the repo for feedback — this keeps the material oriented toward research-backed explanations rather than pure product tutorials.
Who It's For — and Tradeoffs
Great fit if you want a structured, research-oriented introduction to LLMs for classroom use, self-study, or to build a reading list for a research group. The repository is especially useful for educators who need chapter-length units plus curated paper lists.
Look elsewhere if you need hands-on, production deployment tooling (this is a textbook and reading/resource collection, not an inference/serving framework), or if you require the absolute latest benchmarks and model checkpoints packaged as ready-to-run experiments. The project focuses on explanations, pointers to primary literature, and pedagogy rather than turnkey model training pipelines.
Where It Fits
Use this repo as the syllabus backbone for an LLM course, a guided reading list for new researchers joining an LLM team, or a structured reference when you want to connect implementation choices back to foundational papers. For hands-on tutorials (notebooks, training scaffolds, or inference stacks) pair it with practical toolkits from Hugging Face, vLLM, or instructor-provided labs.
