- Jun 9, 2017
- 58
- 277
It can definitely do NSFW stuff but you still need to add like a system prompt/initial prompt saying it's ok to do those things but it doesn't have to be very long (one I've been using uses about 150 tokens (also those 150 tokens is more than just the nsfw prompt, it has some other instructions for the AI to follow to be better at RP, so the nsfw prompt itself is just 32 tokens)). There are no like hard filters unlike on the actual ChatGPT site (which I think you might be able to disable with some browser extensions).I really like the idea. But isn't ChatGPT super censored when it comes to NSFW responses? Or does going through the API have a little more freedom?
Last edited: