Work

Joint spectral edge limits for the Gaussian quadratic SYK model

Year: 2026
Type: preprint
Abstract: We study the Gaussian quadratic Sachdev–Ye–Kitaev (SYK) model on the full Hilbert space and establish joint spectral edge limits at three scales. The centered maximum eigenvalue converges to a Gaussian random variable. The rescaled gaps below the maximum converge, as a point process, to all finite subset sums of the even reflected-sine determinantal process on the nonnegative half-line. At growing ranks, a suitably centered and rescaled deficit field converges to a Gaussian field with explicit covariance and smooth paths on compact positive parameter intervals. These three limiting objects are mutually independent. The proof combines determinant asymptotics, a conditional local limit theorem, and an exponentially weighted finite-kernel comparison to connect the one-particle process with individual ordered many-body levels. We identify the growing-rank centering through its (log k) log log k correction and obtain convergence of all positive absolute moments of the centered maximum, each fixed deficit, and each fixed evaluation of the rescaled deficit field. Declaration of AI use Large language models (LLMs) were used extensively in the development of this work and the preparation of the manuscript, including mathematical exploration, the formulation and refinement of arguments, and manuscript revision. The authors take full responsibility for the mathematical arguments, results, and final content of this paper.
Language: English
Open Access status: green
OpenAlex ID: W7236629360
Cites:
Cited by:
Related to: 10