How to improve the quality of GPT responses? 14 proven prompting techniques measured in a case study
It turns out that there is a great benchmark of what prompting methods to use to get response quality. The recent PromptingGuide.ai case study suggests that it is possible to improve ChatGPT output quality from ~66% by default (meaning only 2 out of 3 responses are good quality answers) to ~92% (9 out of 10 are good quality answers).
Not bad, isn’t it? The difference that smart and thoughtful prompting can create is next to astonishing.
If this topic is interesting to you, vote with 🤯 emoji to the post, and I will analyze the key techniques in one of the following posts.
It turns out that there is a great benchmark of what prompting methods to use to get response quality. The recent PromptingGuide.ai case study suggests that it is possible to improve ChatGPT output quality from ~66% by default (meaning only 2 out of 3 responses are good quality answers) to ~92% (9 out of 10 are good quality answers).
Not bad, isn’t it? The difference that smart and thoughtful prompting can create is next to astonishing.
If this topic is interesting to you, vote with 🤯 emoji to the post, and I will analyze the key techniques in one of the following posts.