4 ms·
You could probably keep the Claude slop hidden and have a fresh model generate a paraphrased response for anything human visible and keep the Claude responses a
by not_paid_by_yt 2mo ago
You could probably keep the Claude slop hidden and have a fresh model generate a paraphrased response for anything human visible and keep the Claude responses as thinking.