I've actually encountered situations with the current gen of "curated" LLM's where legitimate good-actor questions (such as questions around sex or less-orthodo

I've actually encountered situations with the current gen of "curated" LLM's where legitimate good-actor questions (such as questions around sex or less-orthodox relationship styles or wanting a sarcastic character response style, etc.) were basically "nanny-torpedoed", if you know what I mean. To that end, what's the current story with regards to "bare" open-source LLM's that do not have "wholesome bias" baked into them?