Neural networks are already clean. Goodfire's new essay on what is inside AI and why Eliezer Yudkowsky melted down over it.
Why listen
It goes beyond the title with direct discussion of neuron, model, network, including: Goodfire's new essay on what is inside AI and why Eliezer Yudkowsky melted down over it.
Key takeaways
01Interpretability is the science of understanding what is going on inside an AI model
02The ambitious version aims to understand all of the model
03You can ask an AI model a question and read its answer, but you can't see why it gave you that particular answer
Best for
research-minded practitioners comparing model behavior