Shocking to me how little interest the general public has in mechinterp given the alien capabilities demonstrated by LLMs - this and the subsequent transformer-circuits.pub publications will be seen as classic, foundational work in a few years
> A previous project, the Distill Circuits thread, has attempted to reverse engineer vision models, but so far there hasn’t been a comparable project for transformers or language models.
Nothing about B-H curves or magnetization current, must be that other kind of transformer, in that other kind of circuit.
Someone should step up and rectify that, there's a lot of potential for confusion.
This comment is hysteresical
I find it polarizing.-
It can be quite eddyfying
A harmonious comment about an impedance mismatch quickly spiraled into a galvanizing flow
It's just that HN is currently saturated with this kind of thing
Shocking to me how little interest the general public has in mechinterp given the alien capabilities demonstrated by LLMs - this and the subsequent transformer-circuits.pub publications will be seen as classic, foundational work in a few years
> A previous project, the Distill Circuits thread, has attempted to reverse engineer vision models, but so far there hasn’t been a comparable project for transformers or language models.
How successful was that?
The entire thing went down:
- https://distill.pub/2021/distill-hiatus/
... and, here we are.-
I’ve tried reading this many times. But it’s sooooo long. Is it worth it?
Neal Nanda's walkthrough of it is pretty good.
https://youtu.be/KV5gbOmHbjU?is=GSiv0rSHBocNdU8l
Also long, but it's detailed and complicated, so there is no escaping that.
https://en.wikipedia.org/wiki/Royal_Road#A_metaphorical_%22R...
This is a gem.-
Depending on your field, degree of interest, general curiosity, its a must. Its up there with word2vec.-