scalingfactbullishA three-term scaling law that accounts for model size, training steps, and batch size correctly recovers the scaling of optimal batch sizeMachine Learning (Statistics)27 Jul 2026http://arxiv.org/abs/2607.01487v1