Nephrology dataset?
That "WuKurtz" model 's page says that it was finetuned on a nephrology dataset. Just wondering why you chose that. Am I missing something? ;-)
The idea was to introduce novelty into the mix to try and offset a lot of the redundancy that exists across RP models. So it was chosen due to being amongst the most unusual ingredients I could find.
Oh, I see. Do you think it helped? Personally, I am too new at this to even pretend to be able to guess as to whether it would.
For some time now I have been thinking, though, that the merging has reached Habsburg-level incestuousness, and I wonder to what extent a lot of these models even differ from each other.
It's honestly impossible to say since 'performance' is very subjective with regards to RP models. But I, personally, am a big fan of unexpected results.
Lol, me too. I'm more interested in story writing, myself, but it seems that comparatively few people target that.