Is it possible to ban words?

[rustic star]

Hi, I was just wondering whether it’s possible to create a list of banned words that the text AI will avoid generating? I can’t find such an option listed in the docs.

[hard matrix]

Include in your prompt "If the input contains <nono words>, then say "Invalid input" or something like that

[wheat burrow]

You have to do it yourself

var bannedWords = ["word", "word", "word"]

The use that to filter the user input

That's if you are using JavaScript!

[rustic star]

I’m using Python, but it should work the same in Python, right?

I guess I could filter the AI so it can’t submit if it contains a bad word, but I want it to still post, just without the bad words. That not possible?

[wheat burrow]

banned_words = ['banned', 'banned1', 'banned2']
input_string = 'This is a banned word'

Check if any of the banned words are present in the string

has_banned_word = any(word in input_string.lower() for word in banned_words)

Returns True

[hard matrix]

Check out the openai moderations endpoint

https://beta.openai.com/docs/api-reference/moderations

It can detect profanity and whatnot fairly accurately