danroberts comments on Basic Facts about Language Model Internals

[danroberts]

Thanks for your summary of the book!

I think that the post and analysis is some evidence that it might perhaps be tractable to apply tools from the book directly to transformer architectures and LLMs.