WRSWorld Research StateContribute
WORLD RESEARCH STATE

EssentialAI: Muon compute-time frontiers and telescoping tuning

Contribution grade

Optimizer evaluation and hyperparameter transfer · Human-designed research method or resource

Reference score45,000Provisional editorial estimate of this exact contribution; its grade is assessed separately.

What changed

Studied Muon batch-size-dependent compute/time frontiers and combined it with muP using a multiscale telescoping search that narrows expensive large-width sweeps.

Scientific scope

Studied Muon batch-size-dependent compute/time frontiers and combined it with muP using a multiscale telescoping search that narrows expensive large-width sweeps. Inherits Muon, muP, hierarchical search and Moonlight scaling. This is not invention of Muon or guaranteed global Pareto dominance; finite-width/mesh-error analysis assumes smooth asymptotic behavior. Tuning overhead can be substantial, not free.

Evidence

The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. Inherits Muon, muP, hierarchical search and Moonlight scaling. This is not invention of Muon or guaranteed global Pareto dominance; finite-width/mesh-error analysis assumes smooth asymptotic behavior. Tuning overhead can be substantial, not free.

Contributors

Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.

Full scientific record · Browse contribution grades · Request a correction