Logan Riggs comments on Finding Features in Neural Networks with the Empirical NTK