`Then there are the "tortured phrases" that strongly suggest a paper has been substantially written using LLMs. A recent conference paper on statistical methods

`Then there are the "tortured phrases" that strongly suggest a paper has been substantially written using LLMs. A recent conference paper on statistical methods for detecting hate speech on social media produced several, including "Head Component Analysis" rather than "Principal Component Analysis," "gullible Bayes" instead of "naive Bayes," and "irregular backwoods" in place of "random forest."`