Ambiguous out-of-distribution generalization on an algorithmic task 2025-02-13T18:24:36.160Z

The slingshot helps with learning 2024-10-31T23:18:16.762Z

Comment by Wilson Wu (wilson-wu) on Alexander Gietelink Oldenziel's Shortform · 2025-03-03T06:11:12.690Z · LW · GW

"Utter elitism" is a nice article about this phenomenon

Comment by Wilson Wu (wilson-wu) on Ambiguous out-of-distribution generalization on an algorithmic task · 2025-02-16T17:32:48.398Z · LW · GW

For that earlier section, we used smaller models trained on $S_4$ intersect $A_4\times 2$ (4,000 parameters) instead of $S_5$ intersect $A_5\times 2$ (80,000 parameters) -- the only reason for this was to allow for a larger sample size of 10,000 models with our compute budget. All subsequent sections use the $S_5$ models.