4 papers · 1 filter
Hidden Gauge Controls Feature Specialization in ReLU Networks
Tongxi Wang
The success of deep learning depends on learning useful representations, yet predicting how training organizes these representations across neurons remains difficult. In this work,…
Sharp Spectral Thresholds for Logit Fixed Points
Tongxi Wang
Softmax feedback systems are a common mathematical core of entropy-regularized reinforcement learning, logit game dynamics, population choice, and mean-field variational updates. T…
Tracking Drift: Variation-Aware Entropy Scheduling for Non-Stationary Reinforcement Learning
Tongxi Wang, Zhuoyang Xia, Xinran Chen +1
Real-world reinforcement learning often faces environment drift, but most existing methods rely on static entropy coefficients/target entropy, causing over-exploration during stabl…
Stability of In-Context Learning: A Spectral Coverage Perspective
Tongxi Wang, Zhuoyang Xia
In-context learning (ICL) is a pivotal capability for the practical deployment of large-scale language models, yet its reliability can vary substantially with the number of demonst…