ACAR: Adaptive Complexity Routing for Multi-Model Ensembles with Auditable Decision Traces Paper • 2602.21231 • Published Feb 6
Orion: Characterizing and Programming Apple's Neural Engine for LLM Training and Inference Paper • 2603.06728 • Published Mar 6
KALAVAI: Predicting When Independent Specialist Fusion Works -- A Quantitative Model for Post-Hoc Cooperative LLM Training Paper • 2603.22755 • Published Mar 24
