CNWPlayer

hi

Well actually wait, it's more like half and half, half is obvious AI (em dashes, it's not X it's Y, etc) while the other is human generated ("remeaning", "Yet the gap is soo small on benchmarks")

Not that hard to rewrite a singular sentence... or to read your own blog post...

Did ChatGPT write this?! Because, no offense, it looks like a generic GPT-generated... literally anything.

The biggest takeaway isn't that 100B tokens is some magic number.

It's that training-token count alone doesn't determine how good a small language model will be.

But BananaMind 2 Pro has reached the point where the question isn't just:
...

It's becoming:...

"It's not just X, it's Y"

task	Boris-1.7	BM-2-Medium	Δ
HellaSwag	31.22	32.43	−1.21
ARC-Easy	45.88	43.81	+2.07
ARC-Challenge	26.02	25.34	+0.68
PIQA	62.57	61.86	+0.71
4-task average	41.42	40.86	+0.56

Arithmark 3? and bananamind bench?

will be interesting to see the effects of this