[dominotw]

>“Mechanistic interpretability will probably never reduce large language models to a few simple equations,” Icard concluded, “but it may gradually turn deep neural networks into systems whose hidden algorithms can at least partly be understood.”

what is the basis for this optimism ?

[azakai]

The optimism is based on the successes so far, some of which are described in this article. Scientists have made progress here.

[dominotw]

no they havent . success so far is totally meaningless and doesn't imply any sort of upward slope .

[azakai]

The researchers in the field disagree with you. Look at conferences like NeurIPS and ICLR to see a steady stream of incremental progress in this area.

[dominotw]

trust the researchers bro

[Bjartr]

What would real progress look like?