Insanity 2 hours ago

You are absolutely right!

But common patterns of LLMs today will become adopted by humans as we are influenced linguistically by our interactions - which then makes it harder to detect LLM output.

AbrahamParangi 2 hours ago

This is an artifact of RLHF and far better human facsimiles are trivial with uncensored / jailbroken models.