3 papers
cs.LG2026
Unstable Features, Reproducible Subspaces: Understanding Seed Dependence in Sparse Autoencoders
Gleb Gerasimov, Timofei Rusalev, Nikita Balagansky +3
Sparse autoencoders (SAEs) are widely used to interpret neural network representations, but their utility depends on whether the learned features are reproducible across training r…
hep-th2025
Entanglement Entropy in Jackiw-Teitelboim de Sitter gravity with Timelike Boundaries
Timofei Rusalev
The consideration of timelike boundaries in de Sitter static patches has a broad motivation, such as the formulation of a well-defined canonical ensemble and the realization of a n…
hep-th2024
Black Holes, Cavities and Blinking Islands
Dmitry S. Ageev, Irina Ya. Aref'eva, Timofei A. Rusalev
Placing a black hole in a cavity provides a natural framework for exploring gravitational scales, thermodynamic instabilities, and effective gravity theories. In this paper, we exa…