The wide range of text used to train the models allows them to generate output with different tones. ``` Translate the following from slang to a business letter: <Dude, This is Joe, check out this spec on this standing lamp.> ``` > [!GPT-3.5]- > Dear [Recipient's Name], > > I hope this message finds you well. I wanted to bring your attention to an interesting specification regarding the standing lamp we have. > > Best regards, > Joe ``` Translate this text into slang: < I would like to invite you to go to have a drink with me at your earliest convenience. > ``` > [!GPT-3.5]- > Hey, wanna grab a drink sometime? Hit me up when you're free! [[language translation]] < [[Hands-on LLMs]]/[[5 Prompting]] > [[spell and grammar check]]