Produced as part of the SERI ML Alignment Theory Scholars Program - Winter 2022 Cohort Introduction I think that Discovering Latent Knowledge in Language Models Without Supervision (DLK; Burns, Ye, Klein, & Steinhardt, 2022) is a very cool paper – it proposes a way to do unsupervised mind reading[1] –...
Preface We would like to thank the following people who contributed to the generation of ideas and provided feedback on this post: Alexandre Variengien, Daniel Filan, John Wentworth, Jonathan Claybrough, Jörn Stöhler, June Ku, Marius Hobbhahn, and Matt MacDermott. We are a group of four who participate in SERI ML...
Produced as part of the SERI ML Alignment Theory Scholars Program - Winter 2022 Cohort
I think that Discovering Latent Knowledge in Language Models Without Supervision (DLK; Burns, Ye, Klein, & Steinhardt, 2022) is a very cool paper – it proposes a way to do unsupervised mind reading[1] – diminished only by not making its conceptual coolness evident enough in the paper writeup. This is in large part corrected in Collin Burns's more conceptual companion post. I'm rooting for their paper to sprout more research that finds concepts / high-level features in models by searching for their shape[2]. The aim of this post is to present a conceptual framework for this kind of interpretability,... (read 5588 more words →)
We would like to thank the following people who contributed to the generation of ideas and provided feedback on this post: Alexandre Variengien, Daniel Filan, John Wentworth, Jonathan Claybrough, Jörn Stöhler, June Ku, Marius Hobbhahn, and Matt MacDermott.
We are a group of four who participate in SERI ML Alignment Theory Scholars Program under John Wentworth, and we are extending the paper “Discovering Latent Knowledge in Language Models Without Supervision" as we see it as an interesting direction where we could contribute concrete progress towards alignment.
We have a number of potential directions we could explore. The goal of this post is to get feedback on them early in order to prioritize better. This... (read 3435 more words →)
Are there any updates about when the books would be available on Amazon UK or even better Amazon DE / FR to make it easier for people based in EU countries to get them?
Thanks for making the books, I'm really looking forward to get them! :-D
For a 'Bayesian Probability 101", I'm currently following Richard McElreath's Statistical Rethinking course.
I still haven't finished it (only on Chapter 4), but so far it's all that I wanted it to be:
It's funny cause this was recommended to me by an EA friend, so I assumed someone would have mentioned it here already. ;-)
The link from this reply now posts to a steroids page. From the DOI in the link, I found the article here:
https://journals.sagepub.com/doi/abs/10.1111/j.0956-7976.2005.01576.x
(for anyone interested and still looking at these comments 13y later ;-) )
I mistakenly signed twice. Will there be a duplicate check, or could you simply remove my second entry?
[I forgot I had NoScript enabled, so after the first attempt nothing seemed to have happened. That's why I disabled it, which refreshed the site, and submitted again. Then I saw that first time worked already, so now my name appears twice.]
The epilogue reminded me of a funny quote I've heard in a conversation about Computer Security and Encryption:
"If brute force isn't working, you're not using enough"
This is the first meetup in an attempt to create a LessWrong group in Geneva.
The plan for this meetup is to:
See you there!
And following up on this, "Socrateses" is probably wrong. 😅
In Modern Greek, the plural would be Socratides (Σωκράτηδες; the primary stress is on a) or Socrates (Σωκράτες; way less commonly used). With a 2-min search I found this ref to make the case for Socratides.
[And since I happen to have an Ancient Greek language teacher in the next room, by asking her, she gave the following reference]
In Ancient Greek, it would be Σωκράται. If you look at section "133. α" here you can find its conjugation in the example. This would be translated to either: Socrate, or most probably Socratai (again with the primary stress on "a" and with the last "ai" pronounced as the "ai" in "air".
Given the above, the term originally used by Duncan (Socrati), is pretty damn accurate.