Language Models' interpretability

nlp
1

Is there any efficient way to analyze how a piece of knowledge is encoded in LMs?

2

Honestly. Understanding the machanics of Transformers can really help you. Pytorch Transformer encoder