Prompt Engineering

TOPIC: Prompt Engineering

Definition of Prompt Engineering

Prompt Engineering is the skill of designing and refining the best instructions, or "prompts," to get the most useful and accurate results from a Generative AI. It's the art of giving clear and detailed commands to an AI to make it do exactly what you want. A good prompt is like a good recipe; if you give a robot chef a clear recipe, it will make a perfect meal. If you just say "make food," you won't get a good result.

Key Techniques of Prompt Engineering

To write a good prompt, remember to be S.C.A.R.P.E.D.

1. S - Specificity: Be as detailed and precise as possible. A vague prompt will give you a vague answer.
2. C - Context: Give the AI background information. Tell it who you are, what the topic is, and what the purpose of the output is.
3. A - Ask a role: Tell the AI to act like a specific person or expert, such as "Act as a history teacher" or "You are a senior computer programmer."
4. R - Refine and Iterate: If you don't get the result you want, don't give up. Try a new prompt or add more detail to your last one.
5. P - Provide Examples: Show the AI an example of the kind of response you want. This is a very powerful way to get the perfect output.
6. E - Explain steps: Break down complex tasks into a list of clear steps. For example, "First, summarize the article. Second, list the key points. Third, explain the main idea in one sentence."
7. D - Define the format: Tell the AI exactly how you want the answer to be structured, for example, "Write the answer as a bulleted list," or "Present the information in a table."

Examples of Good vs. Bad Prompts

Here is a table that shows the difference between a bad, vague prompt and a good, detailed one:

Bad Prompt	Good Prompt
Write about space.	Act as a space scientist. Explain the importance of planets for a JSS 3 student. Use simple language and include a fun fact.
Write a story.	Write a short, engaging story about a young boy who discovers a hidden AI robot in his backyard. The story should be a maximum of 200 words and end with a surprise.
Summarize a text.	Summarize the following article in three bullet points. Make sure to use simple words: [Insert the article text here].

Test Your Knowledge

Select the correct answer for each question and click 'Submit' to see your score.

1. What is Prompt Engineering?

A) The process of fixing errors in an AI. B) The skill of asking an AI the right questions for a better result. C) The skill of writing computer code. D) The ability of an AI to learn from data.

2. To get a better result from an AI, a good prompt should be:

A) As short as possible. B) Vague and simple. C) Clear and specific. D) Written in a foreign language.

3. Asking an AI to "Act as a professional writer" is an example of which technique?

A) Providing Examples. B) Explaining Steps. C) Asking a role. D) Defining the format.

4. What should you do if an AI gives you a bad result?

A) Stop using the AI. B) Refine your prompt by adding more detail. C) Blame the AI for being wrong. D) Ask the same vague prompt again.

5. Telling an AI to "Write the answer as a bulleted list" is an example of:

A) Providing Context. B) Providing a Specific Example. C) Defining the format. D) Asking a role.