Pretraining Task Diversity and the Emergence of Non-Bayesian In-Context Learning for Regression

[@ml-ok3xq]

41:46 looks like spontaneous symmetry breaking

[@andrewkelley7062]

has anyone else seen the ringing that happens at lower node systems in this situation. The duality problem, and who stole my notes, or maybe great minds think alike. By the way I feel like a 30 % separation works best 49:39 sort of a broad possible boundary.

[@andrewkelley7062]

Sort of a yes no I don't know to be determined situation