Home Artificial Intelligence OpenAI adds custom guidelines to ‘ChatGPT’

OpenAI adds custom guidelines to ‘ChatGPT’

4
OpenAI adds custom guidelines to ‘ChatGPT’

(Photo = Open AI)

OpenAI has added a function that permits you to set an outline suitable for the aim upfront in ‘ChatGPT’. In case you set your personal characteristics or requirements upfront, ChatGPT remembers them and creates a solution appropriate to the situation.

For instance, if a third grade science teacher uses this function, he won’t must re-enter with the prompt ‘I’m teaching third grade science’ each time he uses ChatGPT to show.

OpenAI said on its blog on the twentieth (local time) that it added such a ‘custom instruction’ function to ChatGPT and made it available by choosing the beta function within the settings menu.

To make use of this feature, you might want to set your preferences for ChatGPT’s responses in two different prompt windows before opening the chat window.

In the primary window, users can tell ChatGPT what it must learn about them, and within the second window, users can specify tips on how to reply to prompts. Guidelines are limited to 1500 characters per window.

Prompt window for setting guidelines (Picture=OpenAI)
Prompt window for setting guidelines (Picture=OpenAI)

For instance, if a third-grade science teacher in an elementary school used ChatGPT to create a lesson plan, he needed to enter considerations akin to the age of the coed and the topic to be taught at first.

Example of setting up custom guidelines (Picture=OpenAI)
Example of organising custom guidelines (Picture=OpenAI)

Custom guidelines can set various aspects, akin to setting the language to make use of for writing programming code, the variety of words to make use of for writing documents, and the tone of voice.

OpenAI explained, “Through communication with users in 22 countries, we received feedback that there was an issue within the strategy of starting a latest conversation with ChatGPT.” “Customized guidelines are an added function to resolve these problems, and once set, there isn’t a need for repeated learning or motion by considering the content in all conversations.”

This feature is currently in beta, starting with the Plus plan, and shall be available to all users in the approaching weeks.

Reporter Park Chan cpark@aitimes.com

4 COMMENTS

LEAVE A REPLY

Please enter your comment!
Please enter your name here