OpenAI is rolling out the option of adding custom instructions in your user preferences to tailor ChatGPT responses.
ChatGPT delivers more accurate responses when you give it context before prompting it. Subsequent ChatGPT responses take your context and further explanations into account to further refine the responses. The problem is that it forgets all of that context when you start a new conversation.
ChatGPT will allow users to add custom instructions in their account preferences to provide context to be used in all conversations without having to provide the context at the start of each new conversation.
Fields available in the preferences section include instructions on what you want ChatGPT to know before responding, and how you would like the response to be formatted.
Some examples in OpenAI’s announcement made it clear how useful this new feature will be.
If you are a 3rd grade science teacher you could ask ChatGPT to craft lesson plans without having to remind it each time what grade and subject you taught.
If you’re a programmer and you let ChatGPT know that you only work with Golang then when you ask it for code it won’t default to Python.
Introducing Custom instructions! This feature lets you give ChatGPT any custom requests or context which you’d like applied to every conversation. Custom instructions are currently available to Plus users, and we plan to roll out to all users soon! https://t.co/fVIM9GeYk2
— OpenAI (@OpenAI) July 20, 2023
The custom instruction feature is currently only available in beta for ChatGPT Plus users with plans to roll it out to other users in the next few weeks. The beta opt-in is available on both the web and iOS applications with no word on whether it will be available on the anticipated Android release yet.
OpenAI has recently received criticism of ChatGPT’s performance with reports seeming to show that its coding and math responses have been getting worse over time. While the custom instructions won’t address those concerns, having a more steerable version of ChatGPT will ensure that the responses are more to your liking with less upfront prompting work.