Yeah, I think all those scientific papers miss out a lot by not checking how people actually RP with models, especially in regards to the prompts used.

Didn't read the article, huh?