We’ve trained Claude to flag potential misuse in conversations, a technique we call self-assessment. Right after it responds, Claude gives a preliminary safety assessment of the conversation.
Since we began using self-assessment in June, it has surfaced misuse that our other safeguards weren’t designed to catch.
Because Claude has already read the conversation, self-assessment is very efficient. And unlike classifiers, which target a discrete set of harms, it draws on Claude’s own judgment, so it gets more accurate as our models improve.
Read more on our new Safeguards blog:
https://t.co/18j6zdHub4
@nini_incrypto_ Deeply informative and profoundly thought-provoking. It's mirroring thoughts I've had with things I've seen myself, but never had the words to properly articulate. No matter what path people choose, I hope it is one that they pick for their own selves.
@r27709484 Is that model actually there in Endfield's folders or is it modded in? What I mean to ask is, if you actually turn her shoes and socks invisible through just the game itself, will there be something there?