ML4Good retweeted
I wrote a mech interp reading list two years ago, but a LOT has changed - announcing v2!
This is a highly opinionated list of favourite papers, not a lit review - I try to summarise key takeaways, critique, advise on which parts to prioritise vs skip, and explain WHY I like it!
ML4Good retweeted
Replying to @MATSprogram
@MATSprogram is now hiring for a Research Manager role based in London! Come help us grow the AI safety research field :)
matsprogram.org/careers
ML4Good retweeted
I can strongly recommend applying if you‘re interested in technical AI safety (or not, but you‘re very good at ML and we can try to convince you to work on alignment).
ML4Good camps are awesome:
ML4Good retweeted
If any of you are want to quickly upskill in in AI Safety, and are in either the UK or Germany, I couldn’t recommend these workshops more
ML4Good retweeted
Yoshua Bengio is looking for theory folk to join him to work on Bayesian approaches to AGI safety. I think this is a great opportunity: I've quite enjoyed the theory discussions I've had with Yoshua so far, and would love more work in this direction.
ML4Good retweeted
Very exited to be part of this. Go apply!
Applications to ML4Good Germany 2024 are open!
Dates: Sept 23 - Oct 3
Application deadline: July 14
Interested in upskilling in AI safety, exploring the research landscape, and connecting with like-minded individuals? Consider applying!
ml4good.org/courses/ml4good-…
ML4Good retweeted
Extremely happy to be able to help organise ML4Good Germany this year. We just opened applications!
Applications to ML4Good Germany 2024 are open!
Dates: Sept 23 - Oct 3
Application deadline: July 14
Interested in upskilling in AI safety, exploring the research landscape, and connecting with like-minded individuals? Consider applying!
ml4good.org/courses/ml4good-…
ML4Good retweeted
What a remarkable job of balancing accessibility and technical accuracy! If you're familiar with the content, it's very fun to watch. If you're not familiar, it's a very nice way to dip a toe in.
In our new video, we talk about research on interpreting InceptionV1, a convolutional neural network. Researchers have been able to understand the function of neurons and channels inside the network and uncover visual processing algorithms by looking at the weights. The work on InceptionV1 is early but landmark mechanistic interpretability research, and it functions well as an introduction to the field. We also go into the rationale and goals of the field, and mention some more recent research near the end. Our main source material is the circuits thread (distill.pub/2020/circuits/) in the Distill journal and an article (distill.pub/2017/feature-vis…) on feature visualization. Authors of this research include @ch402 @nickcammarata @ludwigschubert @gabeeegoooh @mpetrov @shancarter @csvoss @zzznah
(video in replies)













