---
title: Debugging and Refining Outputs
units: [Problem Solving With AI]
summary: You can improve AI's responses.
weight: 80
---

{{% param summary %}}

## Today's Objectives
- Apply debugging techniques to improve AI-generated responses.
- Explain how prompt refinement improves AI outputs.

## Lesson Overview

### What is prompt engineering?
{{% define "Prompt Engineering" %}}

Watch on YouTube: [AI’s Hottest New Job Pays Up to $250K a Year. So I Applied. | WSJ](https://www.youtube.com/watch?v=whkge1rEamU)

### What are the key elements of a prompt?
{{< collapse summary="Click here to reveal the answer." >}}

Just remember, TaCoFoRT:

- **Ta**sk: Specific and actionable instructions
- **Co**ntext: Relevant background details
- **Fo**rmat: How the response should be structured
- **R**eference: Additional examples or resources
- **T**one: The style in which the output is given

{{</ collapse >}}

### How do you refine a prompt?
{{< collapse summary="Click here to reveal the answer." >}}

- Test
- Observe the output
- Update your input
- Repeat until you get the desired result

{{</ collapse >}}

## Assignment
{{% instructions-unit-journal-update %}}
{{% unit-journal-define-terms "Prompt Engineering" %}}

### Five Key Elements of a Prompt
Refine the following prompts using each element, then record your revised prompt and how refining it improved the AI's output.

- Task: "Help me clean my room."
- Context: "You are a marketing executive at a sportswear company. Come up with a slogan for a new running sneaker."
- Format: "You are a professional food critic. Generate some ideas for how to evaluate a new restaurant."
- Reference: "You are an architect planning a building downtown. Figure out how many shops can fit on the ground floor of the new building, which is the size of a football field."
- Tone: "You have been selected as a host for an awards ceremony. Prepare your opening speech."

{{% instructions-code-org-update %}}
{{% unit-journal-question-of-the-day question="How can you improve AI's responses?" %}}