Loading TutorKit...
Prompting
The durable fundamentals of prompting a model well: stating the goal first, supplying the context it lacks, using constraints and examples deliberately, specifying the output precisely, and debugging a weak prompt like an engineer instead of guessing at wording.
Questions in this set
- What is a prompt?
- Why should the goal come first?
- What context should a prompt include?
- How do constraints improve an answer?
- When do examples help?
- How should output requirements be stated?
- Does assigning a role help?
- How do you break down a complex request?
- How do you debug a weak prompt?
- What prompt mistakes recur?
- When is prompting not the solution?
TutorWant me to explain it differently?
AI concepts can be dense. Tell me what's confusing and I'll find a new analogy.