- Joined
- Apr 7, 2019
Jailbreaks. OpenAI actively monitors to prune any workarounds people are dumb enough to publicly post, but if you work with one often and/or get a good understanding of how LLM logic works, alongside a bit of creative writing, you can trick a censored LLM into giving you nsfw results.How do these coomers even get there? Anytime I try to have chatgpt talk about anything adult or controversial it shuts down and tells me it can't.
