HAPPY BIRTHDAY DANIEL
#2
by StentorLabs - opened
HAPPY BIRTHDAY DANIEL!!!
Did you really think I'd let this benchmark pass quietly?
I know the kind of model card you vibe with.
The kind that reports clear scores no matter what.
So I kept reading.
And reading.
And reading.
At the first landing:
BLiMP: 64.96%
At the next landing:
BLiMP: 75.94%
Daniel.
I know the kind of evaluation you vibe with.
The kind that keeps climbing no matter how many percentages stand in its way.
But even I can't climb past two different BLiMP scores.
Why aren't the staircases the same?
Damn. Nice work. That whole msg probally took you 30 minutes.
Anyway, it's fixed up now. The staircases are now indeed the same and you can aura monster your way up them now. Just please, no "kill, kill, kill, kill."
🤍🤗
Harley-ml changed discussion status to closed