Nathan Helm-Burger comments on Sparse trinary weighted RNNs as a path to better language model interpretability

[Nathan Helm-Burger]

I don’t know, but I’d love to know! If you find out, please tell me!