Prompt Engineering: Efficiency in the Age of AI
Chain of Thought (CoT)
Chain of Thought (CoT) is a strategy for guiding an AI model through a series of prompts or queries to guide its responses in a particular way. This strategy is especially handy when you want to dive deeply into a topic or keep the model on a certain path of thinking. Let's see some examples:
When asked about the 5th word in "A B C D", ChatGPT might respond with "D" as the 5th word, but this is not the correct answer.
Note: ChatGPT3.5 is used in this example.
However, by using the CoT technique, you can guide the model to understand the context and respond correctly. Here's an example of the prompt we will use:
User: What is the 5th word in "A B C D"?
Assistant:
- The phrase is "A B C D".
- We count each word sequentially:
- "A" is the 1st word, "B" is the 2nd word, "C" is the 3rd word, and "D" is the 4th word.
- The 5th word is the one after "D", but there isn't one in this phrase.
- So, there is no 5th word in "A B C D".
User: What is the 7th word in "X C Z A L"?
Assistant:
- The phrase is "X C Z A L".
- We count each word sequentially:
- "X" is the 1st word, "C" is the 2nd word, "Z" is the 3rd word, "A" is the 4th word, and "L" is the 5th word.
- The 6th word is the one after "L", but there isn't one in this phrase.
- So, there is no 7th word in "X C Z A L".
User: What is the 7th word in "Q W E R T Y"?
Generative AI For The Rest Of US
Your Future, DecodedEnroll now to unlock all content and receive all future updates for free.
Hurry! This limited time offer ends in:
To redeem this offer, copy the coupon code below and apply it at checkout: