Reaction
An AI benchmark framed as a test of recall
A commenter says current-generation models are good at precision but poor at recall in their hands-on experience.
TLDR
In a reply, a commenter argues that the benchmark is better framed as a test of recall. Based on their ad hoc, hands-on experience, they say current-generation models are good at precision but not good at recall.
Combined views
247
1 Source, first seen ago