The Life Cycle of a Massive Activation: Stochastic Birth, Weight-Decay-Driven Growth, and Competitive Consolidation Paper • 2610.00423 • Published about 1 month ago • 6
Hidden Dynamics of Massive Activations in Transformer Training Paper • 2508.03616 • Published Aug 5, 2025 • 19
Hidden Dynamics of Massive Activations in Transformer Training Paper • 2508.03616 • Published Aug 5, 2025 • 19