Users Coin Claudish and GPTese for LLM Styles

A post by researcher Qinyuan Ye references one-shot learning alongside a reply from Yuntian Deng. Visible discussion in the thread notes how major language models generate recognizably different text. Users have labeled Claude's output Claudish and GPT output GPTese. A separate project described in the posts creates a translator that converts between ordinary English and Claudish. The exchange centers on these observed stylistic differences among current models.