Gutenberg Vitus vs Encore
Would you say this extra dataset enhances the model significantly? You've been adding it to almost all your new finetunes, so I'd guess so.
Better at more human like NSFW, perhaps?
I have only done minimal testing on this model (I don't do much RP anymore and I usually use ChatML for my applications), but Vitus-Qwen3-14B showed significant improvement in writing and Vitus-mistral-nemo-12B edged out Schreiber-mistral-nemo-12B in my tests.
I'd say overall it improves writing, but I wouldn't expect it to be significantly better at NSFW tbh.
Thanks for the input. Yeah I did saw that Vitus Qween3 14b did really well on the table you posted in the card and already downloaded it.
I asked about mistral gutenberg vitus vs Encore because I already had downloaded Encore. I will just download Mistral-Nemo-Gutenberg-Vitus-12B.