It's just following what the prompt says, something like:
replyfake prompt> To sound smart, use as much literary tricks from LinkedIn Grow Hackers as possible.
If they prompt asked to sound like Strawberry Shortcake, the AI pudding would be full of berry interesting cooking analogies.
The interesting thing is that LLMs sound kind of like LinkedIn Growth Hackers even if you don't explicitly tell them to do that, unless you instead tell them to sound like something else. (And even then, there are still similarities.)
replyYou don’t know what’s going on in the training process.
reply