Improve ChatGPT results using a series of three prompts.
This "chaining" technique has shown to reduce the number of errors ChatGPT makes by 50% in some cases.
This is the equivalent to raising a student's grade from an 86 to a 94!
AI Explained's work is impressive, thorough, and traceable back to his tests. We were not, of course, able to inspect all of his tests. But, we will be testing this technique as we optimize the work we do.
In fact, his sequence solves one of the problems highlighted in the recent TED Talk, Why AI Is Incredibly Smart — and Shockingly Stupid | Yejin Choi
This is not perfect, but it does show how to use reflection to improve results without outside input.
Try the prompting framework below to see if it improves the results of your technical and logical GPT requests.
Question. <Enter your question>. Answer: Let's work this out in a step by step way to be sure we have the right answer.
You are a researcher tasked with investigating the <X> response options provided. List the flaws and faulty logic of each answer option. Let's work this out in a step by step way to be sure we have all the errors: <Copy your options from Prompt 1>
You are a resolver tasked with 1) finding which of the <X> answer options the researcher thought was best 2) improving that answer, and 3) Printing the improved answer in full. Let's work this out in a step by step way to be sure we have the right answer.
<Copy your options from Prompt 2>