Asked about Working out what the network is doing
What can interpretability actually establish today?
An answer that works
Specific circuits and specific behaviours have been explained, mostly in smaller models. Explaining an arbitrary output from a frontier model is not yet possible. It is real progress on a problem that is far from solved.
Reading an answer is not being able to give one. The section that teaches this is Working out what the network is doing, 4 hours of Interpretability.
Learn this properlyAlso asked about Interpretability
What is mechanistic interpretability?What is superposition and why does it make interpretability hard?What is the difference between probing and activation patching?What is superposition and what is the response to it?Why is attribution not an explanation?What is the residual stream, and why is it a useful way to think about a transformer?Walk me through an activation patching experiment.How would you check whether an SAE feature really means what its label says?
All 10 questions on Interpretability