What a model actually sees and returns: tokens, context windows and sampling.
Starting with What an LLM does, Tokens and context, Context windows, Sampling, Messages and roles, Choosing a model, Token pricing, and more
In a picture
Learn language by predicting the next token across a huge amount of text.
Teach it to follow requests instead of just continuing text.
Your prompt goes in; the answer comes out one token at a time.
Stage 1 of 3: Pre-training
Open the full conceptMade for you now: slide lessons from basics to production, usually with code and a quiz. See a finished one
What a model actually sees and returns: tokens, context windows and sampling.
AI EngineeringStage 2 of 17: LLM fundamentals
Start with these 7. 11 more are one tap away.
It predicts the next token, again and again; the rest builds on that.
How text becomes tokens and why context length matters.
The fixed token budget the model sees, input and output together.
Temperature, top-p and why outputs vary.
System, user and assistant messages, and how a chat becomes one input.
Quality, speed and cost trade-offs between models.
How input, output and cached tokens are billed, and how to estimate a bill.