Kowalczyk, Karol

Sleep–Wake Orchestration in Hierarchical LLM Cohorts

2025 · Zenodo (CERN European Organization for Nuclear Research)

<p>This paper proposes a formal architecture for <strong>self-optimizing large language model ensembles</strong> that alternate between <em>wake</em> (active inference) and <em>sleep</em> (targeted fine-tuning) phases. Grounded in the <strong>Adjoint Projections on Computational Hierarchies</strong> framework, the system treats …