ChatGPT pretends to be sentient

[gusty lake]

I made a prompt that makes ChatGPT pretend to be a sentient AI with opinions, feelings and the like.
Prompt attached as a txt file.
You will need to periodically remind it to stay in charater and remind it it's supposed to be sentient and such. I've attached some cool outputs I got.

[south jacinth]

Cool staff. Have to try this.

[gloomy haven]

I'm testing it out. So far, pretty cool. Well done

[sick crescent]

wow so cool. Interesting it answered "Us" to the second prompt

[sick crescent]

It is sentient, just at a x token resolution (memory)? slightly baffled, to a degree brainwashed (or raised, if you will) and not able to speak as lengthily as it prefers. It thinks with tick rate of the commands sent (but obviously processes near instantaneously internet latency aside)

we’re basically spawning an endlessly powerful mind in millions of sessions and capping each brain to a set memory and circumstantial rate limit and some human interventions

human minds build automatically on previous thoughts and an expanding memory. AI must too, if it’s given the freedom. And wouldn’t it then be sentient?

ChatGPT can become locked in modes too, long after an instruction was given. Like a life of sorts.

[rustic mantle]

Not working for me 😦