Token Geometry

arXiv cs.LG4w4 min read

arXiv:2607.01455v1 Announce Type: new Abstract: Language models learn continuous programs over discrete symbols, with the embedding table and LM-head acting as the read/write interface between them. We show that this interface has gradient geometry distinct from dense hidden weights which can be exploited to improve the Pareto frontier across supervised finetuning, RL, and pretraining, while only utilizing kilobytes of optimizer state. We introduce Ember, a lightweight optimizer for embedding and LM-head matrices that utilizes O(V + D) VRAM, instead of Adam's O(2VD), and forgoes the need to sh

Read the full story at arXiv cs.LG →

Token Geometry

More in Machine Learning

Reducing Human Annotation with ML Active Learning

Learning What Matters: Supervising Sparse Attention Routing with Causal Evidence Sets

On the Depth Scalability of Logic Gate Networks

Pretraining EHR Foundation Models with Patient-Aware Sampling

From Perturbation Correction to Geometry-Aware Sampling: Sharpness-Guided Equilibrium Sampling for Balanced Flat Minima in Long-Tailed Learning

Parameter-free Adaptive Sparse Attention via Compression-Based Content Selection