I have noticed changes in ChatGPT based on month+ long on-going "discussions" with it that brush up against its content policy and ethics guidelines. Are these changes learnt, or actively programmed?
I am not wanting to delve into the details behind what I have been working on with CHATGPT (it's not illegal), or how I have been able to trial-and-error my way around convincing it to provide responses to requests it signals as potentially violati…