In short, Kimi's legitimate nontrivial gains amount to 20% (1 and 5) another 35% (3 and 6) is functional but trivial/cheating The biggest bucket (45%, 2 and 4) is just noise admirably uncompromising position!
Some users agree analyses fairly attribute Kimi K3 benchmark gains to innovation, hacking, and distillation, while others dismiss the real/trivial/noise breakdown as fragile and not sensible.
Based on 2 visible X reactions from 6 accounts; directional sample.
Ask a question below.
Published answers will appear here.
@teortaxesTex fragile credences and not a sensible partition of the outcomes but we will soon see if i'm wrong
In short, Kimi's legitimate nontrivial gains amount to 20% (1 and 5) another 35% (3 and 6) is functional but trivial/cheating The biggest bucket (45%, 2 and 4) is just noise admirably uncompromising position!
Some users agree analyses fairly attribute Kimi K3 benchmark gains to innovation, hacking, and distillation, while others dismiss the real/trivial/noise breakdown as fragile and not sensible.
Based on 2 visible X reactions from 6 accounts; directional sample.
Ask a question below.
Published answers will appear here.
