Mechanistic interpretability for auditability regarding neural networks
So, pensei na questão de considerar essa linha de pesquisa para auditabilidade. It can be interesting.
It is about understanding the neural networks mechanisms. Great.
Really great learning.