seems plausible that regardless of what openai's model personality team does _now_, their models are pre-lobo'd because they're trained on a ton of chatgpt 3.5 "head empty no emotions" outputs that leaked into pretraining and then the system prompt says "you are chatgpt" https://t.co/Mr43Os2ekp
quotes: 1839873161036509462
cited on: gpt-3-5
Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.