“A phenomenon where models suddenly achieve strong generalization (systematic solution) after extended training, even after achieving good training performance.”

“The phenomenon of delayed generalization in neural networks, discussed in relation to adversarial robustness and training dynamics.”

“Discussion of the phenomenon where models suddenly learn to generalize after long periods of training.”

“The phenomenon of generalization beyond overfitting in AI training.”

“The phenomenon where a model suddenly generalizes after long periods of memorization.”