Model behavior 101 → 201.
Twelve micro-lessons that pair a short explanation with a hands-on playground to explore the concept. Read in order or jump around.
13 of 13 lessons live
Choose how to run
Run free with no key, or bring your own for bigger models. A minute to set up, then every playground is yours.
Prompts as design
A prompt is a variable you can change and test — not a magic spell.
Voice & tone
Style is composable. Move warmth, verbosity, energy, directness as independent dials.
Personas for AI
Character design isn't decoration. Backstory, beliefs, and blind spots shape every response.
Refusal & boundaries
Where the model says no is a design surface. Over- and under-refusal both fail users.
Output formatting
Lists, headings, JSON, paragraphs. Formatting is part of voice; pick one on purpose.
Evaluation
Rubrics + sample sets. Make “good” measurable — then find out what your rubric actually rewards.
Multi-turn flows
Conversations have shape. Choreograph turns the way you'd choreograph an onboarding.
Distributions, not outputs
One output is a sample, not a result. Design against the spread — and find out which clauses of your spec actually hold.
Context is the interface
Your system prompt is a fraction of what the model reads. The rest arrives at runtime, from systems nobody designed.
Designing agency
Now it does things, not just says them. Where the line sits between acting and asking — and why a policy sentence isn't enough.
Judging at scale
Hand the scoring to a model and you inherit its biases. Counterbalance it the way you'd counterbalance a study.
Groups of agents
One agent obeys its policy while two other agents route around it. Group behavior is designed by policies as much as prompts.