Gryphe/Codex-24B-Small-3.2 · This model is fantastic!

I've been using this as Gryphe_Codex-24B-Small-3.2-Q6_K_L. I've ran this through all sorts of different scenarios with different tones in Silly Tavern and it's delivered amazing results every time. It was really good with the suggested settings applied over the ST defaults but with some tweaking it became even better. The training data was clearly quite diverse as it almost doesn't matter what off the wall subject I come up with - it knows what I'm talking about. Its ability to understand context, subtext, sarcasm, and jokes is probably the best that I've seen so far. The lack of comments on this model is almost criminal! Keep up the good work.