From Latent Space to Training Data: Explainable Specialization in Minimal MLPs 事件

From Latent Space to Training Data: Explainable Specialization in Minimal MLPs arXiv:2605.25939v1 Announce Type: cross Abstract: We here study whether training biases can make hidden neurons specialize in minimal one-hidden-layer MLPs, and whether such specialization improves prototype-based reconstruction of the training dataset from the learned weights. We consider Gaussianactivation MLPs of width equal to dataset size and compare three structural losses that respectively encourage coverage o