Advanced · Advanced · 6 rounds · ~7 min · free with an account
Multi-turn Refinement
Conversation as compute. Use 3-5 turn loops when the goal is discovered, not pre-specified.
Prompt Arcade · 6 rounds · free with an account
Multi-turn Refinement
Use the conversation as a workbench: diverge, choose, refine, then lock the format. Each turn should carry one clear change.
- Order it
- Spot it
- Build it
- Match it
- Spot it
- Build it
Five games are open to everyone — start with Prompting Mindset 101. The other 31 need a free account (no card) so your scores and progress are saved.
What you'll learn
- The refinement arc: Diverge → choose → deepen → critique → format.
- Turns that stall: Every turn should contain a change the model can make.
- Critique turn: Critique from a specific viewpoint, then fix only what it found.
- Turn types: Know which kind of turn you're taking.
- Context drift: When early rules fade, restart with a summary of decisions.
- The restart summary: Restart with decisions, not transcripts.
More advanced games
Self-Consistency VotingRun the same prompt 3-5 times → take the majority answer. +12-18% accuracy on reasoning tasks (Wang et al. 2022).Few-Shot Pattern RecognitionShow 2-4 examples → the model maps the relational structure, not the surface details (Chase & Simon's chess masters, applied).Constraint CalibrationToo few constraints → drift. Too many → brittle. The 5-constraint sweet spot and how to find it.