🌱 AI Engineering · How LLMs Work (Visual First) · cozy lesson
What are LLMs?
10 min · 1 min read · no scary math, promise
Big autocomplete, carefully trained
An LLM predicts the next token (word piece) given previous tokens. Repeat that loop and you get sentences, code, plans.
Training in one line: pretrain on lots of text → teach instructions → align with human feedback.
Why they sound confident when wrong
They are trained to continue fluently, not to say “I don’t know”. Without evidence in the prompt or tools, gaps become fluent guesses — hallucinations.
Fixes you will learn: RAG grounding, abstain rules (“say not found”), citations, evals.
Tokens matter
“ChatGPT is great” ≈ 4-5 tokens. Costs and limits are in tokens, not words. That’s why chunking and top-K matter for RAG cost.
Next
Tokenization → Transformer simplified → your first API call.
Check your understanding
Correct answers earn XP (once each).
1. LLMs generate text by…
My notes (saved in this browser)
Select text above → Save selection, or write your own. AlgoMaster-style notebook, local-first for MVP.
No notes yet. Your highlights will live here.
Finished reading? Seal it with a tick ✅
The checkbox in the explorer turns green too — same progress.