10 - Large Language Models

This module covers LLM core technologies from Transformer architecture to practical applications.

Module Structure

10-large-language-models/
├── 01-llm-fundamentals/        # LLM Basics
├── 02-pretrained-models/       # Pretrained Models
├── 03-fine-tuning/             # Fine-tuning
├── 04-prompt-engineering/      # Prompt Engineering
├── 05-rag/                     # RAG
├── 06-agents/                  # Agent Systems
└── 07-alignment/               # Alignment Training

Core Content

01 - LLM Fundamentals

Transformer architecture
Tokenization (BPE, WordPiece)
Pretraining objectives

02 - Pretrained Models

GPT series: Autoregressive generation
BERT: Bidirectional encoding
LLaMA: Efficient open-source

03 - Fine-tuning

Full fine-tuning
LoRA: Low-rank adaptation
QLoRA: Quantized LoRA

04 - Prompt Engineering

Zero-shot / Few-shot
Chain-of-Thought
ReAct

05 - RAG

Document chunking
Vector embeddings
Vector databases (FAISS, Chroma)

06 - Agent Systems

Tool calling
Memory management
Task planning

07 - Alignment

RLHF
DPO