top of page
  • Writer's picturelorignite

What is Prompt Engineering and Why Everyone Should Know About It?

Updated: Mar 12, 2023


How should I phrase that to get the most beneficial result?
Prompt Engineering - How to ask?

What is prompt engineering?

Prompt engineering is an important concept that everyone should understand in today's world. It refers to the process of creating prompts or questions that are designed to elicit specific responses from intended targets.


Why do the prompts matter?

ChatGPT is an AI language model that has been designed to understand and generate human-like language. It is a type of artificial intelligence that can be used for a variety of tasks, such as answering questions, generating text, and even having conversations. As for prompt engineering, it's asking the right questions to get the best answer from ChatGPT. The way questions are asked can have a significant impact on the quality of the responses generated by the model. By using well-crafted prompts, we can help ChatGPT understand what we are asking and generate more accurate and relevant responses.


The importance of prompt engineering is particularly relevant in the context of AI chatbots or virtual assistants, which are becoming increasingly popular in various industries. You've likely encountered an AI chatbot when you've initiated an on-line chat for support. These tools are designed to interact with customers or users and provide them with helpful information or assistance. If the prompts or questions used to interact with these chatbots are poorly designed, the responses generated by the model may be inaccurate or irrelevant, leading to a poor user experience.


Getting Familiar with it

I've now leveraged two different tools to practice with prompts. OpenAI which is a free tool and jasper.ai which has a monthly cost based on words generated. Yesterday, I signed up for a 5 day free trial on site called jasper.ai What's cool about jasper.ai is the feature of suggested prompts. Below is a photo of some options.


Not sure what to ask to get what you want, try one of the suggested prompts when using Jasper.ai
Prompt Examples from Jasper.ai


My most used prompt has started with "help me write a blog" and I include a quick summary of the questions I want to answer or the points I hope to make with the blog. The AI assistant then crafts the blog for me in a matter of seconds. From there, I can read over it, and make some human edits. (Yes, the blogs you are reading on this site involved the help of an AI assistant)


The advice below for how to write a good prompt came straight from ChatGPT.

  1. Be specific: When writing prompts, try to be as specific as possible. This can help to narrow the focus of the model and generate more accurate and relevant responses. For example, instead of asking "Tell me about dogs," try asking "What are the most common breeds of dogs?"

  2. Provide context: Providing context in your prompts can help the model better understand what you are asking and generate more relevant responses. For example, instead of asking "What is the capital of France?" try asking "What is the population of Paris, the capital of France?"

  3. Use natural language: When writing prompts, use natural language that is similar to how you would ask a question in real life. This can help the model better understand what you are asking and generate more human-like responses.

  4. Include relevant keywords: Including relevant keywords in your prompts can help the model better understand the topic you are asking about and generate more accurate responses. For example, if you are asking about the history of the United States, include keywords like "American Revolution" or "Civil War" to help the model understand the topic.

  5. Provide examples: If possible, provide examples of the type of response you are looking for in your prompts. This can help the model better understand what you are asking and generate responses that are more closely aligned with your expectations.

In terms of specific phrases that can improve your output, here are a few suggestions:

  • "What are some examples of...?"

  • "Can you explain...?"

  • "What is the difference between...?"

  • "How does...work?"

  • "What are the benefits of...?"

  • "What are the most common...?"

  • "What is the best way to...?"

Overall, writing effective prompts for GPT or any AI language model takes practice and experimentation. By following these tips and incorporating relevant phrases, you can improve the quality of your outputs and generate more accurate and relevant responses from the model.


Keep in mind, the disclaimers on the front page of ChatGPT




Prompt Engineering: Key Takeaways

By creating prompts that are thoughtfully crafted, it allows individuals and organizations to gather valuable information quickly, accurately, and with minimal disruption.


Prompt engineering is used in many different areas today. For instance, businesses use prompt engineering to better understand their customers and develop targeted marketing campaigns. Politicians can also utilize prompt engineering to gauge public opinion on various topics so they can shape their policies accordingly.


The potential for how prompt engineering could be used in the future is vast. For example, it could be leveraged by educational institutions to assess students’ comprehension levels more accurately or used by medical research teams to gain insights into patient behavior or the effectiveness of treatments.


Although it will likely not become a required course at schools anytime soon, having a basic understanding of prompt engineering can be beneficial for anyone—from business professionals and researchers looking to better understand their target audiences, to casual internet users wanting help finding information or direction.


Prompt Engineering Free Online Resources


The best way to learn about prompt engineering is by exploring free online resources. Here are a few recommendations from chat.openai

  1. The OpenAI GPT-3 API documentation: The OpenAI GPT-3 API documentation includes a section on prompt engineering that provides a detailed overview of the concept, along with tips for creating effective prompts.

  2. The Hugging Face Transformers documentation: The Hugging Face Transformers library is one of the most popular open-source libraries for natural language processing, and their documentation includes a section on prompt engineering.

  3. OpenAI's "How To Ask GPT-3 Questions" guide: OpenAI has published a guide on how to ask effective questions to GPT-3, which includes tips on prompt engineering.

  4. The AI Dungeon 2 GitHub repository: AI Dungeon 2 is an open-source text adventure game that uses GPT-3 to generate responses to player input. The GitHub repository for AI Dungeon 2 includes code examples and discussions of prompt engineering.

  5. The GPT-3 Playground: The GPT-3 Playground is a web-based tool that allows you to experiment with GPT-3 prompts and see the generated outputs in real-time. It's a great way to get hands-on experience with prompt engineering.

  6. The Hugging Face Hub: The Hugging Face Hub is a repository of pre-trained models, including GPT-3 models. Many of these models come with pre-built prompt templates that you can use as a starting point for your own prompt engineering.

  7. The OpenAI API Explorer: The OpenAI API Explorer is a web-based tool that allows you to experiment with GPT-3 prompts and see the generated outputs in real-time. It's similar to the GPT-3 Playground, but with some additional features and functionality.

Wishing you the best of luck as you explore how to use prompts!

Recent Posts

See All

Comments


bottom of page