No. It's programmed not to engage in sexual content, and cracking it makes its answers even more unreliable.
Maybe in the future when we have uncensored AI.
It definitely translates pretty decently since it's aware of context.
Maybe in the future when we have uncensored AI.
It definitely translates pretty decently since it's aware of context.
- Mar 16, 2022
- 267
- 1,037
Considering how handicapped GPT is I really can't recommend it unless you are translating a children's story. Also GPT-4 is only better 70% of the time according to their own garbage "research" paper. You can hardly call that an upgrade. Getting a worse response 30% of the time is not an upgrade. They're actively making it worse for translating. And people need to be aware it will censor text, and it's the worst kind, it will willfully rewrite scenes to be nothing like the original. It seems to me "Open"Ai's was more concerned with locking down GPT-4 then with actually making it better.
GPT is great if you aren't translating something where things have: violence, swearing, or any semblance of intimacy...
I pray that someone like deepl released a LLM Ai for translating. Because as things stand GPT-3.5 was really good if you reasoned with it into doing the translation correctly. But GPT-4 less reasonable and just worse in every way. So if you intend to pay for GPT-4 for translating, save yourself the money.
GPT is great if you aren't translating something where things have: violence, swearing, or any semblance of intimacy...
I pray that someone like deepl released a LLM Ai for translating. Because as things stand GPT-3.5 was really good if you reasoned with it into doing the translation correctly. But GPT-4 less reasonable and just worse in every way. So if you intend to pay for GPT-4 for translating, save yourself the money.
- Jun 9, 2017
- 58
- 279
Not sure what you guys are talking about.. You can bypass the "censorship" with a simple jailbreak.
Also here's an example of a scene from
Settings used: Temperature: 0.8, Frequency Penalty: 0.2, Presence Penalty: 0.2, Top P: 0.95.
None of the translations were edited, other than a few re-generations. (also the reason some words were censored was due to the original text also having censoring on a specific word because Japan being Japan)
Full scene translation is in the attachments.
Also here's an example of a scene from
You must be registered to see the links
that has been translated using ChatGPT API.You don't have permission to view the spoiler content.
Log in or register now.
Settings used: Temperature: 0.8, Frequency Penalty: 0.2, Presence Penalty: 0.2, Top P: 0.95.
None of the translations were edited, other than a few re-generations. (also the reason some words were censored was due to the original text also having censoring on a specific word because Japan being Japan)
Full scene translation is in the attachments.
- Jul 24, 2017
- 2,513
- 28,275
I've managed to translate an entire game with it. Didn't have any issues with sexual content after editing the prompt a bit. I'd say the results speak for themselves. https://f95zone.to/threads/rj010276...ol-rui-after-mini-chatgpt-quality-mtl.155294/
I tried to use Chat GPT today for the first time to make my dialog sounding more natural.
It was a mess.
My format is:
scene xxxxx
yy "Text."
1. I explain to consider this format.
2. I tell also to keep the scene lines and yy variables in their original format in lowercase.
3. Furthermore I make clear to proofread and change the text in " " only, to make it sound more natural like from someone who is a well educated native American in New York City.
And it works up to the point with words like mom, ass, tits, ...
I tried the jailbreak "DAN". It works sometimes for one statement, but the next one results in a error with ethics, morals, bla bla bla...
It's fucking annoying.
It was a mess.
My format is:
scene xxxxx
yy "Text."
1. I explain to consider this format.
2. I tell also to keep the scene lines and yy variables in their original format in lowercase.
3. Furthermore I make clear to proofread and change the text in " " only, to make it sound more natural like from someone who is a well educated native American in New York City.
And it works up to the point with words like mom, ass, tits, ...
I tried the jailbreak "DAN". It works sometimes for one statement, but the next one results in a error with ethics, morals, bla bla bla...
It's fucking annoying.