Good Morning!
Delivering interesting content every single week on Web3, Security, Crypto, NFTs, Design & AI.
It's FREE, Takes less than 5-minutes to read, and you are guaranteed to learn something.
Subscribe to get valuable Web3 News, Useful Resources and Insights every week to your Inbox!
ChatGPT is a large language model developed by OpenAI, based on the GPT-3.5 architecture. One of the unique features of ChatGPT is its ability to prompt users to generate a chain of related thoughts or ideas in response to a particular question or topic. This technique, known as chain of thought prompting, is a powerful tool for promoting cognitive flexibility and adaptive thinking.
Chain of thought prompting in ChatGPT involves prompting the user to generate a chain of related thoughts in response to a particular question or topic. The model then uses these responses to generate a more complex and nuanced response, based on the user's initial thoughts and ideas. This process allows users to explore the relationships between different ideas and gain a deeper understanding of their thought processes.
To illustrate the technique, consider the following example: A user might ask ChatGPT, "What are the benefits of regular exercise?" ChatGPT might then prompt the user to generate a chain of related thoughts in response, such as:
Exercise helps to improve cardiovascular health.
It can reduce the risk of chronic diseases such as diabetes and heart disease.
Regular exercise can improve mood and reduce anxiety and depression.
Exercise can improve sleep quality and reduce insomnia.
It can help to maintain a healthy weight and improve body composition.
Exercise can improve cognitive function and reduce the risk of cognitive decline.
Based on these responses, ChatGPT can then generate a more complex and nuanced response that takes into account the user's initial thoughts and ideas.
How to Use ChatGPT?
Chain of thought prompting in ChatGPT can be used in a variety of contexts, including education, therapy, and decision-making. In education, it can be used to help students develop their critical thinking skills and to encourage them to explore complex concepts and ideas. In therapy, it can be used to help individuals overcome negative thought patterns and to develop more adaptive beliefs and attitudes. In decision-making, it can be used to help individuals explore the potential consequences of different options and to make more informed and rational decisions.
There are several different techniques that can be used to facilitate chain of thought prompting in ChatGPT, including keyword prompts, open-ended prompts, and guided prompts. Keyword prompts involve prompting the user to generate a chain of related thoughts based on a specific keyword or phrase. Open-ended prompts involve prompting the user to generate a chain of related thoughts based on a more general question or topic. Guided prompts involve providing the user with a set of specific questions or prompts to guide their thinking and exploration of a particular topic.
ThinkGPT - implementing Chain of Thoughts for LLMs. Try here.
Chain of thought prompting in ChatGPT has several benefits. It allows users to explore complex topics and ideas in a structured and organized manner, promoting cognitive flexibility and adaptive thinking. It also allows users to gain a deeper understanding of their thought processes and to identify and challenge negative or irrational beliefs.
However, there are also some limitations to the technique. One potential limitation is that ChatGPT may not always generate a response that accurately reflects the user's initial thoughts and ideas. Additionally, the technique may not be suitable for individuals with severe cognitive impairments or those who are experiencing acute distress or anxiety.
Chain of thought prompting is a powerful tool for promoting cognitive flexibility and adaptive thinking, and it is a key feature of ChatGPT. The technique involves prompting the user to generate a chain of related thoughts in response to a particular question or topic, which can then be used to generate a more complex and nuanced response. While there are some limitations to the technique, overall, it has several benefits and can be used in a variety of contexts, including education, therapy, and decision-making.
Few examples of chain of thought prompts that ChatGPT could generate:
If you could have any superpower, what would it be and why? What are some ways you could use this power to help others?
What are some benefits and drawbacks of social media? How has social media changed the way we communicate and interact with one another?
What are some potential solutions to climate change? How can we balance the need for economic growth with the need to protect the environment?
What are some of the causes of income inequality? How does income inequality affect different groups of people, and what can we do to address it?
What are some of the ethical implications of artificial intelligence? How can we ensure that AI is developed and used in ways that benefit society as a whole?
These prompts are designed to encourage individuals to generate a chain of related thoughts and explore complex ideas in a structured and organized manner. By generating a chain of related thoughts, individuals can gain a deeper understanding of a particular topic and develop their critical thinking skills.
One potential application of chain of thought prompting in ChatGPT is in the field of education. By prompting students to generate a chain of related thoughts in response to a particular question or topic, teachers can encourage them to explore complex concepts and ideas in a structured and organized manner. This can help students to develop their critical thinking skills and to gain a deeper understanding of the subject matter.
Another potential application is in the field of therapy. Chain of thought prompting can be used to help individuals overcome negative thought patterns and to develop more adaptive beliefs and attitudes. For example, if a person is struggling with anxiety or depression, ChatGPT can be used to prompt them to generate a chain of related thoughts about their thoughts and feelings, which can then be used to identify and challenge negative beliefs and to develop more positive and adaptive attitudes.
Chain of thought prompting can also be used in decision-making. By prompting individuals to explore the potential consequences of different options, ChatGPT can help them to make more informed and rational decisions. This can be particularly useful in situations where there are multiple options available and the consequences of each option are complex and difficult to predict.
Overall, chain of thought prompting is a powerful tool that can be used in a variety of contexts to promote cognitive flexibility and adaptive thinking. While there are some limitations to the technique, overall, it has several benefits and is a key feature of ChatGPT. By prompting users to generate a chain of related thoughts in response to a particular question or topic, ChatGPT can help individuals to explore complex ideas and gain a deeper understanding of their thought processes.
Programming in prose
For slightly more advanced prompts, think about what you are doing as programming in English prose. You can give the AI instructions and it mostly-sort-of follows them. Mostly, because there is a lot of randomness associated with AI outputs, so you will not get the consistency of a standard computer program. But it can be worth thinking about how you can provide a very clear and logical prompt to the AI.
A lot of active research is happening around the best way to “program” an LLM, but one practical implication is that it can help to give the AI explicit instructions that go step-by-step through what you want. One approach, called Chain of Thought prompting, gives the AI an example of how you want it to reason before you make your request, as you can see in the illustration from the paper.
You can also provide step-by-step instructions that build on each other, making it easier to both check the output of each step (letting you refine the prompt later), and which will tend to make the output of your prompts more accurate. Here is an example: ChatGPT is generally really bad at creating interesting puzzles and scenarios to solve, either making things too easy or impossible. But if we are explicit and step-by-step about what we want, the results are much better:
You a game master. Your job is to come up with interesting challenges for the player to solve. Describe a challenging fantasy scenario, and enable me to solve it in an interesting way. You will use the following format to help create a series of responses.
Chain of thought:
[Step 1]: Decide on the the scenario, making it original and vivid and not standard fantasy. The scenario can involve combat, a trap, or a puzzle. The scenario must not involve riddles or the elements. Make sure there is a solution to the scenario. Make the solution require clever thinking. Include the solution in [] brackets
[Step 2]: Decide on the scene. Make sure that the player has the option to solve the scenario based on the descriptions. Make sure the solution is not clear, but requires clever reasoning based on the scene. Make sure there are very different false solutions that seem plausible. Include the detailed true solution and describe the false solutions, as well as how the player would find the true solution in [] brackets
[Step 3]: Describe the scenario and the scene, vividly and originally. Make sure there are clues to the solution and credible, but very different, false clues to the wrong answer in the description. Do not describe the solution or the problem directly. Do not describe how to solve the problem in this step. Do not describe the false clues as false.
Begin by introducing yourself and go through each step in order."
Again, there is no single template here. You need to experiment with your “program” multiple times to get a good answer, but there is an art to it that you can learn with time.
A little bit of magic
Okay, I lied slightly about the “there-are-no-magic-words” thing. There are some phrases that seem to work universally across LLMs to provide better or different results, again, by changing the context of the answer. Some possible things to experiment with:
Be creative/make any assumptions you need. This will tend to remove some of the constraints of practicality around AI answers, and can be useful if you are trying to generate something novel.
Show your work/provide sources/go step-by-step. The AI will make up information that it does not have access to. There is some evidence that asking it to show its work, or its sources, reduces that risk somewhat. Even if it doesn’t, it can make checking work easier.
Write me code and tell me how to use it. If you can’t code, you might be able to now. AI can do some amazing things with Python programs, and tell you exactly how to run it. I don’t know coding, but I have written a dozen Python programs in the last month. If there are errors in the code, and there likely will be, just give them to the AI to correct.
Write a draft/provide an example. If the AI refuses to do something (“you should be creative and write your own novel, I can’t help”, sometimes asking it to provide something like a draft can get it to produce results.
The secret is practice.
Prompting is a skill that can be learned and improved with practice. The best way to learn is to use AI as your partner and your teacher. Experiment with different types of prompts and see how the AI responds. Engage in a dialogue with the AI and ask it questions, give it feedback, and challenge it to do better. Find out what triggers the AI’s creativity and what limits it. You will soon develop your own style and technique that will make you a more effective prompter. And remember, the AI is always learning too, so keep up with the latest developments and innovations in the field. You never know what new possibilities you might discover. One more thing: don’t be discouraged if some prompts don’t work as expected. There is some randomness in the AI’s output, and sometimes you will need to use the “clear” button to start a new conversation. That’s part of the learning process too.
If you're enjoying today's newsletter, why not share it with your friends? They might find it just as informative and entertaining as you do.
Sharing is caring, and by spreading the word about this newsletter, you're helping to support ME and ensure that more great content gets produced in the future. Plus, you'll get to have even more conversations with your friends about the interesting topics covered in each edition.
So go ahead and hit that share button.
I hope this was helpful!
Thank you for reading!
If you're interested in following along, feel free to subscribe!
Let’s bust some more in next article.
If you want more, be sure to
