Do we have runs with the old models but the newer harness?
I see that someone else has already asked this, oops. It would be nice to have more systematic evaluations of which harness changes help or don’t, perhaps on a simpler task.
Do we have runs with the old models but the newer harness?
I see that someone else has already asked this, oops. It would be nice to have more systematic evaluations of which harness changes help or don’t, perhaps on a simpler task.