How LLMs Work
Tokens, embeddings, attention and next-token prediction — the mechanics that explain almost every LLM behaviour.
- 4 lessons
- 1h 4m
- 3 simulations
Hallucination, prompt injection and guardrails — the risks to understand before an LLM app meets users.
A suggestion, never a lock — you can start here.
Every lesson's title and summary is public. The lessons themselves open once you own the track.
Every lesson in LLMs, Prompt Engineering & Context EngineeringA model picks the most plausible token, not the truest one. The cure is not asking it to "not make things up" but putting the source in context, requesting checkable quotes, and allowing "not found".
The model reads your instructions and a product review through the same channel: tokens. When data contains commands, no prompt can separate them — what can is an architecture that limits the damage.
A guardrail is code that checks inputs and outputs without trusting the model. Monitoring is what tells you quality has dropped — before users do.
Tokens, embeddings, attention and next-token prediction — the mechanics that explain almost every LLM behaviour.
Clear instructions, examples, step-by-step reasoning and structured output — proven techniques, no incantations.
What goes into the context window, where it comes from and in what order — the core skill of building LLM applications.
Calling APIs, streaming, tool calling, caching, and keeping cost and latency under control in production.