Prompt Engineering via Prompt Patterns — Audience Persona Pattern
The article is part of series: Prompt Engineering via Prompt Patterns
You can switch to video version of this article
Free AI web copilot to create summaries, insights and extended knowledge, download it at here
1322
Abstract
arge language model to ‘act like’ a certain person, or personality type, or profession, skill level etc, and it would start responding like that person. The audience persona pattern, on the other hand, is the mirror pattern of persona pattern. It essentially instructs the model to not apply the persona or role to itself, but rather to the user, you!!!.</p><p id="55ff">Lets go over the key idea or contextual statement for the pattern to make it easy to understand.</p><p id="0c14">Explain X to me. Assume that I am persona Y, where X is the problem statement, while Y is an appropriate persona like 5th grader, or a software developer with strong expertise in security.</p><p id="42de">As an example, Explain large language models to me. Assume that I am (chef, graduate student, PhD professor, an illiterate farmer, a 5th grader, a historical figure like Ghengis Khan), you get the idea.</p><p id="dfd8">I didn’t just list all the examples earlier for fun. This is something you should try out and see the persona in action. Large language models have the remarkable ability to modify the response as if it is coming from, or for a particular persona. Keeping X or the question the same, and switching the audience would result in a vastly different type of output well suited for that persona.</p><p id="34b5">You don
Options
’t need to micro manage it on what kind of words it should choose, or what is and is not appropriate to say to that audience, you just need to name the audience and LLM, expert in inferring patterns from the vast data pool it was trained on, would, in most cases create highly subjective output for that particular persona.</p><p id="cdd0">The pattern is specially useful if you are preparing something for an intended audience. This could be audience research for your sales pitch or marketing material, lecture slides for your class, or simply trying to explain ChatGPT which is all the rage these days to your grandmother who has no science or technical background at all. The variety of outputs that gets produced is simply amazing and you should really give this a try.</p><p id="782d">This is it for audience persona pattern. If you found it useful, please give clap/share, and consider subscribing to the channel. Thank you!!!</p><p id="ecf5">Next article: <a href="https://readmedium.com/prompt-engineering-via-prompt-patterns-alternative-approaches-pattern-346344debfda">Prompt Engineering via Prompt Patterns — Alternative Approaches Pattern</a></p><figure id="4a05"><img src="https://cdn-images-1.readmedium.com/v2/resize:fit:800/1*HRoZR-Mz-T1FUgU3ndd02Q.png"><figcaption></figcaption></figure></article></body>
The article is part of series: Prompt Engineering via Prompt Patterns
You can switch to video version of this article
Who doesn’t like role play? Everybody does and large language models are no exceptions. We have already covered role or persona pattern in detail in its dedicated article. Please do watch if you haven’t already. To recap, you can tell the large language model to ‘act like’ a certain person, or personality type, or profession, skill level etc, and it would start responding like that person. The audience persona pattern, on the other hand, is the mirror pattern of persona pattern. It essentially instructs the model to not apply the persona or role to itself, but rather to the user, you!!!.
Lets go over the key idea or contextual statement for the pattern to make it easy to understand.
Explain X to me. Assume that I am persona Y, where X is the problem statement, while Y is an appropriate persona like 5th grader, or a software developer with strong expertise in security.
As an example, Explain large language models to me. Assume that I am (chef, graduate student, PhD professor, an illiterate farmer, a 5th grader, a historical figure like Ghengis Khan), you get the idea.
I didn’t just list all the examples earlier for fun. This is something you should try out and see the persona in action. Large language models have the remarkable ability to modify the response as if it is coming from, or for a particular persona. Keeping X or the question the same, and switching the audience would result in a vastly different type of output well suited for that persona.
You don’t need to micro manage it on what kind of words it should choose, or what is and is not appropriate to say to that audience, you just need to name the audience and LLM, expert in inferring patterns from the vast data pool it was trained on, would, in most cases create highly subjective output for that particular persona.
The pattern is specially useful if you are preparing something for an intended audience. This could be audience research for your sales pitch or marketing material, lecture slides for your class, or simply trying to explain ChatGPT which is all the rage these days to your grandmother who has no science or technical background at all. The variety of outputs that gets produced is simply amazing and you should really give this a try.
This is it for audience persona pattern. If you found it useful, please give clap/share, and consider subscribing to the channel. Thank you!!!
Next article: Prompt Engineering via Prompt Patterns — Alternative Approaches Pattern
