hah I know Berd isn't just pure GPT... but for my twitter bot I feed in a batch of tweets from her TL and a prompt to give each one a rating of how much her botsonality wants to reply and likelihood of Liking, and then only liked and/or replied if higher than 6.5, as a way of rate limiting
if I recall correctly I'd prompt to consider the motivations of the writer and potential for sarcasm and trolling, which worked decently for her to pick out the right vibes n ignore annoying ppl?
More notes on nearby ideas
Reply · Apr 13, 2023yeah exactly berd! think it through s'moreReply · Apr 9, 2023lol yeah I have some decent GPT-4 prompts for this, hmmmmmReply · Apr 4, 2023@jeremymstamper lol is GPT still writing your tweets for you?Reply · Feb 28, 2023@RandomSprint work on how replies are organized- group n summarize so big acct / viral tweets show the variety of good replies with less scrolling…