“ChatGPT Prompt Engineering For developers” is a course that was co-created by deeplearning.ai and OpenAI
It’s an interesting and practical course using Python and OpenAI API (with the API key) to practice the different must-know things about prompt engineering.
It’s a short format and free. You can take it easily.
While it’s essentially built for beginners, the advanced can also find some useful information in it.
Here is the link:
In the following, you will find details on what it talks about:
The Best practice to know to write efficient prompts
- Write clear and detailed prompts
- Use delimiters to contour the text you want to work on
- Ask for structured outputs: JSON, HTML formats…
- Ask the model to check whether conditions are satisfied
- Give sample to chatGPT to allow answer in the same style
Iterative way to generate prompts
- Start with a prompt for a given task
- As you get the answers back from ChatGPT, you can add more information to fine tune your target answer:
- Shorten the answer
- Focus on some details more than others
- Get the answer on a given format: Json, HTML table
- Summarize a text, by giving some instructions:
- Number of words, or characters
- Focus on some topics on the text
- Extract information instead of summarize
- Identifying if the sentiment of a given text/review is positive or negative
- Asking to list the different emotions in a review
- Extracting topics from a given text
- Determining if a given topic is mentioned in a text
- Tone of a discussion
- Spell and Grammar check
- Creating an email answer in a given tone with customization
- Using Temperature to vary the answer details.
- Create a chatbot by injecting back the answers coming from ChatGPT to constitute a context so ChatGPT can remember information.
Here is the link again: Click here
Summary of “ChatGPT Prompt Engineering For developers”
“ChatGPT Prompt Engineering For developers” course equips you with essential techniques to efficiently utilize OpenAI API, enabling you to maximize your productivity.
Using only one API provided by ChatGPT, you can perform a wide range of NLP tasks such as summarization, translation, sentiment analysis, topic modeling, question-answering, chatbot, and formatting without the need to create your own model by gathering, preparing, training, testing, and fine-tuning the parameters. ChatGPT is a potent tool that can help you achieve greater automation efficiency.