Curvature-Conditioned Multiscale Momentum with Sphere Constraints for LLM Pretraining

arXiv:2608.28442 · cs.LG · Submitted 2026-08-28 · Read on arXiv

cs.LG

Submitted: 2026-08-28

Updated: 2026-08-28

Comments: 50 pages

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Related papers