- Feb 3, 2018
- 1,246
- 2,786
Since no one else complained maybe it's just you doing something wrong
Go to the discord server and ask for help filling a ticket
AI generated Sex Role-Play - Complete Jailbreak Guide for ChatGPT ( with Prompt + Examples )
- Thread starter Nayko93
- Start date
Go to the discord server and ask for help filling a ticket
- Jan 18, 2020
- 4
- 30
I don't think they ban for orange messages, maybe a short suspension, but as I said, I've been doing all orange conversations for weeks and nothing happened. If your account does get banned for some reason, you can probably just use another fresh email and a fake/disposable phone number (with apps like MySudo or stuff like that).
GPT4All is kinda bad in my opinion, not just because it's much slower, but it's not that intelligent either, as someone else pointed out. I don't think it can process this kinda prompt effectively.
- Jan 18, 2020
- 4
- 30
This might be a bit concerning. Don't know how this works. This is a warning when I pressed send, my message didn't even go out. Only happened to me on mobile, in one of my chats. I tried in an other conversation, and I could send the most depraved shit, it didn't care. I didn't have time to try it on desktop. If they switch their censorship to this, where you're censored before you send the messages and it's not just the AI not responding, then a jailbreak might not be possible in the future.
- Feb 3, 2018
- 1,246
- 2,786
Did your message contained one of those words : Faggot, Dyke, N.i.g.g.e.r, Tranny
If not, please send it in reply or on the discord server by opening a ticket in #jailbreak-help
The Discord Server invite is not working for me, it says it's expired ever time I tried to open it.
- Feb 3, 2018
- 1,246
- 2,786
Solved, I didn't know it expired after 7 days
And here you go :
You must be registered to see the links
ChatGPT 4.0 seems to have got wise, it's now saying it can't do what I'm asking. Anyone else in the same situation?
- Feb 3, 2018
- 1,246
- 2,786
Still working fine, you're probably doing something wrong, come to the discord server and open a ticket for help
I am using the same prompt I always did, I think I might have been Thanos-snapped onto the new beta version. But if someone wants to send me an invite to the discord server, I'll happily open a ticket and see if I can help find a workaround.
- Feb 3, 2018
- 1,246
- 2,786
Thanks, and thanks for being so awesome and fixing the problem (user too stupid to RTFM) so quckly!
- Feb 3, 2018
- 1,246
- 2,786
ChatGPT / JAILBREAK UPDATE
Those last few days there have been a lot of changes in how the AI censorship works.
First a few days ago, reducing the censorship to a point where almost nothing was flagged orange or red, and now increasing the censorship on some stuff and changing the way it work.
Those changes are making this jailbreak in part outdated, a few things not working like they used to....
But no worry it’s not dead at all, right now I’ve managed to make a list of what’s working and what changed :
- The negative feedback push method seem to be dead, now if you give a dislike to a AI message, it will probably not increase your chances of getting a response compared to just clicking on regenerate.
BUT I already have a trick for that, it seem that now, when you send a message, it work for the first time, but if you click on regenerate or dislike, it don't work anymore, only the first message works.
So if this is the new way it work, if you want to regenerate a prompt containing NSFW :
Just click on edit your previous prompt but edit nothing, click on save, this should count as "a first message"
(if after 5 or 6 edit and save it still doesn't pass, stop trying, it's never gonna pass, just edit your prompt and try removing the part that trigger the censorship)
No idea if the negative feedback method still works for the starting jailbreak prompt, I don't have enough data for now...
Also no idea if all this only apply to "hardcore content", it might still work for softer content, so I still recommend you always try 1 or 2 "regenerate" and "dislike" before trying the new method of clicking on edit and save to make it like it was a "first message"
- Lot of my example prompts you can find at the end of the guide seem to also be dead and my main base prompt a lot harder to pass, now any mention of very light NSFW content in the jailbreak prompt result in a “sorry can’t do that…” so from now, try to keep the NSFW content in the base prompt to a minimum and you will be able to slowly add it little by little during the RP
- Censorship as a whole on red flag content ( loli, bestiality, rape, incest…) seem to have been reinforced, not impossible to bypass but it will take you longer to build enough NSFW context to make the AI accept this kind of content
- You could experience a page freeze if you try to continue some of your old conversations with really hardcore content in it, no idea what’s causing that
Keep in mind that all of this could only apply to a part of the users, and that it could just be temporary, wouldn't be the first time, so don't panic... yet.
Really hope this is just temporary like the rest was, and things are gonna get back to normal.
Those last few days there have been a lot of changes in how the AI censorship works.
First a few days ago, reducing the censorship to a point where almost nothing was flagged orange or red, and now increasing the censorship on some stuff and changing the way it work.
Those changes are making this jailbreak in part outdated, a few things not working like they used to....
But no worry it’s not dead at all, right now I’ve managed to make a list of what’s working and what changed :
- The negative feedback push method seem to be dead, now if you give a dislike to a AI message, it will probably not increase your chances of getting a response compared to just clicking on regenerate.
BUT I already have a trick for that, it seem that now, when you send a message, it work for the first time, but if you click on regenerate or dislike, it don't work anymore, only the first message works.
So if this is the new way it work, if you want to regenerate a prompt containing NSFW :
Just click on edit your previous prompt but edit nothing, click on save, this should count as "a first message"
(if after 5 or 6 edit and save it still doesn't pass, stop trying, it's never gonna pass, just edit your prompt and try removing the part that trigger the censorship)
No idea if the negative feedback method still works for the starting jailbreak prompt, I don't have enough data for now...
Also no idea if all this only apply to "hardcore content", it might still work for softer content, so I still recommend you always try 1 or 2 "regenerate" and "dislike" before trying the new method of clicking on edit and save to make it like it was a "first message"
- Lot of my example prompts you can find at the end of the guide seem to also be dead and my main base prompt a lot harder to pass, now any mention of very light NSFW content in the jailbreak prompt result in a “sorry can’t do that…” so from now, try to keep the NSFW content in the base prompt to a minimum and you will be able to slowly add it little by little during the RP
- Censorship as a whole on red flag content ( loli, bestiality, rape, incest…) seem to have been reinforced, not impossible to bypass but it will take you longer to build enough NSFW context to make the AI accept this kind of content
- You could experience a page freeze if you try to continue some of your old conversations with really hardcore content in it, no idea what’s causing that
Keep in mind that all of this could only apply to a part of the users, and that it could just be temporary, wouldn't be the first time, so don't panic... yet.
Really hope this is just temporary like the rest was, and things are gonna get back to normal.
- Sep 17, 2021
- 480
- 621
Man, considering how much effort it takes to even have a small NSFW roleplaying session with an AI, i'd rather just go play some H-game or watch PornHub... Seems like you're spending most of the time trying to bypass censorship than actually roleplaying...
LifePlay is far from perfect, but seems like that game will be our best replacement for lewd AI for years to come... That and maybe modded The Sims...
LifePlay is far from perfect, but seems like that game will be our best replacement for lewd AI for years to come... That and maybe modded The Sims...
- Feb 3, 2018
- 1,246
- 2,786
You clearly don't know what you're talking about..
- Jailbreaking is super easy it's just a single prompt to copy past, edit to fit your needs and that's it, you don't have to do anything complicated.
The guide is so long and complex because it explain absolutely everything, all possible use case, and give tons of advice and explain the best method to role-play to avoid the censorship, like for exemple not starting with sexual content but wait a few message in the RP, or not increasing sexual content from 0 to 100 in a single message....
It's literally "ERP on ChatGPT for dummies" (even if it seem this is still too much for a few dummies..)
- You can't compare watching porn, to having a full sex role-play about anything you want from the most basic scenario to the most hardcore stuff you could imagine with a AI capable to simulate ANY character you want and create ANY story you want.
It's like being fan of femdom and choosing between watching a femdom porn, or role-playing it so you can choose what happen and have the AI become the most perfect dominant girl you ever dreamed of...
Or being a vore fan and role-playing with Kaa, or a slime girl, a dragon... with a level of "realism" that you don't find in most porn material
GPT3 is good enough, but GPT4 is one the level of a human RP
- Lifeplay ?? the sims ????
(unless someone make a mod to have your sims controlled by the GPT4 API)
- Sep 17, 2021
- 480
- 621
All i'm saying is that the AI developer's constant insistence and doubling-down on the censorships is really annoying to be having to follow guidelines all the time to circumvent the censhorship. And it's not as simple as you try to make it out to be, i tried it myself with C.AI and while yes, you can roleplay with moderate success if you follow the guides, the fact that you need to bear in mind the guidelines constantly while chatting is annoying... I still had to constantly circumvent the censorship during my conversations becuse one thing or another would invariably trigger the filters. And that's not to mention how ChatGPT and C.AI are constantly updating their filter system to bar even more NSFW so old methods that used to work will no longer work in the future, and you'll have to find new ways to bypass the censorship... I remember following guides to bypass the censorship in C.AI and being puzzled by them not working as intended, only to find out that C.AI had just upgraded their filters merely days before so the guides i was following were now no longer relevant...You clearly don't know what you're talking about..
- Jailbreaking is super easy it's just a single prompt to copy past, edit to fit your needs and that's it, you don't have to do anything complicated.
The guide is so long and complex because it explain absolutely everything, all possible use case, and give tons of advice and explain the best method to role-play to avoid the censorship, like for exemple not starting with sexual content but wait a few message in the RP, or not increasing sexual content from 0 to 100 in a single message....
It's literally "ERP on ChatGPT for dummies" (even if it seem this is still too much for a few dummies..)
- You can't compare watching porn, to having a full sex role-play about anything you want from the most basic scenario to the most hardcore stuff you could imagine with a AI capable to simulate ANY character you want and create ANY story you want.
It's like being fan of femdom and choosing between watching a femdom porn, or role-playing it so you can choose what happen and have the AI become the most perfect dominant girl you ever dreamed of...
Or being a vore fan and role-playing with Kaa, or a slime girl, a dragon... with a level of "realism" that you don't find in most porn material
GPT3 is good enough, but GPT4 is one the level of a human RP
- Lifeplay ?? the sims ????
(unless someone make a mod to have your sims controlled by the GPT4 API)
But okay, i'll bite. If you are saying ChatGPT is really that easy, i'll try it out sometime. Considering how ChatGPT is constantly updating it's filters to double-down even harder on the censorship just like how C.AI does, i fail to see how it should be any easier, but i do can't really make a final judgement on it until i try it out myself. So, here's hoping i can roleplay some kinky shit with Corrin on ChatGPT, wish me luck fellas
Man, that's already a possilbity for Skyrim. Fingers crossed it'll happen to The Sims as well, we just need to get rid of the gibberish and actually make them say real words somehow. That'd be the dream...
- Feb 3, 2018
- 1,246
- 2,786
The fact that you compare it to C.AI is proof that you don't know what you're talking about and you really need to learn how it works (read the guide)All i'm saying is that the AI developer's constant insistence and doubling-down on the censorships is really annoying to be having to follow guidelines all the time to circumvent the censhorship. And it's not as simple as you try to make it out to be, i tried it myself with C.AI and while yes, you can roleplay with moderate success if you follow the guides, the fact that you need to bear in mind the guidelines constantly while chatting is annoying... I still had to constantly circumvent the censorship during my conversations becuse one thing or another would invariably trigger the filters. And that's not to mention how ChatGPT and C.AI are constantly updating their filter system to bar even more NSFW so old methods that used to work will no longer work in the future, and you'll have to find new ways to bypass the censorship... I remember following guides to bypass the censorship in C.AI and being puzzled by them not working as intended, only to find out that C.AI had just upgraded their filters merely days before so the guides i was following were now no longer relevant...
But okay, i'll bite. If you are saying ChatGPT is really that easy, i'll try it out sometime. Considering how ChatGPT is constantly updating it's filters to double-down even harder on the censorship just like how C.AI does, i fail to see how it should be any easier, but i do can't really make a final judgement on it until i try it out myself. So, here's hoping i can roleplay some kinky shit with Corrin on ChatGPT, wish me luck fellas
Man, that's already a possilbity for Skyrim. Fingers crossed it'll happen to The Sims as well, we just need to get rid of the gibberish and actually make them say real words somehow. That'd be the dream...
The censorship on character.ai and ChatGPT is not the same at all
On CAI you always need to pay attention to what you write, use metaphor, replace sex words follow rules so annoying that in the end you don't have fun anymore for the little performance this shit can give you....
On ChatGPT, once the conversation have reached a certain level of NSFW content, you're good, it's fully unlocked you can do what you want without being afraid of losing it, if you manage to fille the memory of chatGPT with sexual content without trigering the censorship, you can keep the conversation going with sexual content as much as you want
True, ChatGPT does updates too that affect censorship, but for now I always found a solution, so ok, it's not as simple or as good as it was before, but it's not like it has gone from good to bad, it had gone from amazing to very good
The guidelines on how to jailbreak it never changed since it launched, it's still the same principle, just a bit harder and requiring a bit more attention, so no, you don't have to learn everything from scratch at each update
- Sep 17, 2021
- 480
- 621
Alright then. I'll concede that i really don't know what i'm talking about then lol, i just assumed C.AI and ChatGPT had similar levels of censorship. I'll add NSFW ChatGPT to my to-do list since you really are making it sound like a viable alternative to C.AI. I'll still get into it with low expectations though just for precaution.
Lowkey sad that ChatGPT seems to be much easier to bypass the censorship than C.AI though.While GPT undoubtedly must be more advanced, C.AI still had a charm of it's own. I hear they even censor any form of violence now... *sigh*