Learnable Multipliers: Freeing the Scale of Language Model Matrix Layers
Paper • 2601.04890 • Published • 44
Optimization
-
-
MSign: An Optimizer Preventing Training Instability in Large Language Models via Stable Rank Restoration
Paper • 2602.01734 • Published • 34 -
On Surprising Effectiveness of Masking Updates in Adaptive Optimizers
Paper • 2602.15322 • Published • 11 -
Flash-KMeans: Fast and Memory-Efficient Exact K-Means
Paper • 2603.09229 • Published • 85 -
Why Muon Outperforms Adam: A Curvature Perspective
Paper • 2606.04662 • Published • 10 -
Dion3: Full-Stack Orthogonal Updates
Paper • 2608.11612 • Published • 10 -
Ask Before You Optimize: Dynamic Pre-Formulation Clarification for Interactive Optimization
Paper • 2609.05258 • Published • 18