DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data Paper • 2608.13517 • Published 12 days ago • 32
FlexMoRE: A Flexible Mixture of Rank-heterogeneous Experts for Efficient Federatedly-trained Large Language Models Paper • 2602.08818 • Published Feb 9 • 3
DeToNATION: Decoupled Torch Network-Aware Training on Interlinked Online Nodes Paper • 2502.06728 • Published Feb 10, 2025