If I went to generate paragraphs of text, using several example paragraphs as a source, could anybody help me understand the best approach right now? Let’s use horoscopes as an example: say, I want it to give several (10-20) examples of horoscopes, and then be able ask for “a horoscope for a lonely person”?
Is there something I need to train a model for, or at least a LORA or embedding? Or is that something I can simply pass the examples as context for each time I want to generate a new one?
Thanks for any advice or guidance….