3 ms·
> when you're using the API directly there's no censorship. Wait seriously?
by rcpt 3y ago
> when you're using the API directly there's no censorship.
Wait seriously?
- ineedasername 3y agoThere’s a bit. When I used the openai playground I have received warnings about response potentially being bad, but using the API directly I don’t even get warnings like that. Testing things out, it will produce vile and hateful content on demand. However it won’t say anything. If I specifically tell it to use some words I get the same type of content warning but also an extra note about those words, and that I have to contact openai support if my use case truly requires their use. There’s also the fact that using or distributing such content is against TOS, so I suppose could simply ban you
- reissbaker 3y agoNot exactly. The "censorship" is the RLHF tuning in the chat models as far as I understand it; the API for the chat models is the same AFAIK. The base models don't have censorship, but there are no base models available for API access for GPT-4, only a chat model. You can use the GPT-3-era base models, but, well, they're not as good as GPT-4.
- cubefox 3y agoOr you go to Microsoft Azure and use the GPT-3.5 base model: code-davinci-002. Though they could still use "observational" censorship there, i.e. analyze your prompt with a different model. OpenAI does that, not sure about Microsoft.