Architecture ▸ Transformer layers ▸ Self-attention heads ▸ Feed-forward network Concepts ▸ Tokens & vocabulary ▸ Embeddings ▸ Context window Models ▸ GPT-4o / o3 ▸ Claude / Gemini ▸ Llama / Mistral Applications ▸ Coding assistants ▸ Document analysis ▸ Search & RAG Deployment ▸ Cloud API ▸ Self-hosted ▸ Edge / on-device Limitations ▸ Hallucination ▸ Knowledge cutoff ▸ Bias & cost Training ▸ Pre-training ▸ Fine-tuning (SFT) ▸ RLHF / DPO Techniques ▸ Prompt engineering ▸ Chain-of-thought ▸ Agentic loops LLM ECOSYSTEM
Branches: Architecture Concepts Models Applications Deployment Limitations Training Techniques