id stringlengths 44 67 | source stringclasses 6
values | source_id stringlengths 16 77 | text stringlengths 200 5.07M | num_tokens int64 38 1.3M |
|---|---|---|---|---|
lesswrong::e3dd9435-1689-5ef8-993a-a1aa50bfb186 | lesswrong | y5fhowZqyanr4cDg6 | Can we create a function that provably predicts the optimization power of intelligences?
Follow up to Efficient Cross-domain Optimization
When I am skeptical that we will ever understand intelligence, I am skeptical that we will ever be able to reliably map a systems description onto its optimization power. This has ... | 2,175 |
ea_forum::0a68c575-6873-56fd-a76b-8a78e691021f | ea_forum | Re4iEouCJCHtrzng8 | Reflection on the EA Young Professionals Group
See the original proposal here.
Note: take this data as anecdotal, or at best a case study. There are too little data and too many confounding variables to draw anything more than that out of this pilot—that being said, it hopefully provides something useful.
Summary
T... | 1,295 |
lesswrong::05f1be65-b767-54f8-807d-ed9f3e8f9278 | lesswrong | 3EQQpB7i2o3nu2Di6 | A Bunch of Matryoshka SAEs
This work was done as part of MATS 7.0.
MATS provides a generous compute stipend, and towards the end of the program we found we had some unspent compute. To let this not go to waste, we trained batch topk Matryoshka SAEs on all residual stream layers of Gemma-2-2b, Gemma-2-9b, and Gemma-3-... | 6,204 |
lesswrong::71919807-94be-5706-8022-662eb8e692ff | lesswrong | iYFuZo9BMvr6GgMs5 | Case Study: Interpreting, Manipulating, and Controlling CLIP With Sparse Autoencoders
Click here to open a live research preview where you can try interventions using this SAE.
This is a follow-up to a previous post on finding interpretable and steerable features in CLIP.
Motivation
Modern image diffusion models of... | 3,945 |
lesswrong::113489d4-459e-5f02-bd23-13369a2e5ecf | lesswrong | u4hAr8R82a4HTqkkW | The rare, deadly virus lurking in the Southwest US, and the bigger picture
If you live in this one tiny county in California, you might be more likely to die from Sin Nombre Virus than in a car crash.
In the same way that “why does the frozen spinach I want to buy cost much more than it used to?” engages with a vast ... | 6,192 |
lesswrong::2d517485-9d34-5877-8285-d2cae7b4f178 | lesswrong | ENEYeeyPXyuEQdtZy | Logical Optimizers
EDIT: This idea is unsafe unless you have X and only X optimization. If an early logical optimiser produces a paperclip maximiser, then the paperclip maximiser will play along giving good solutions to most logical optimisation problems, but hiding nasty surprises where we are likely to find them. In... | 1,339 |
lesswrong::7ea5f137-5b77-5268-9342-54d92ceb9d3e | lesswrong | 5bd75cc58225bf0670375502 | Reward learning summary
A putative new idea for AI control; index here.
I've been posting a lot on value/reward learning recently, and, as usual, the process of posting (and some feedback) means that those posts are partially superseded already - and some of them are overly complex.
So here I'll try and briefly summ... | 759 |
lesswrong::07274d97-dbf4-5724-95ff-654ff70782c0 | lesswrong | vc4SSzhiKtHKvFmFA | Factored Cognition with Reflection
Note: I accidentally moved this post to draft and re-published it. My intention wasn't to get it back on the home page. Sorry.
Some weeks ago I completed an HCH-like program/Q&A system that supports reflection, including time travel: https://github.com/rmoehn/jursey I'm just posting... | 105 |
lesswrong::84fd7501-b8f6-5d36-a92b-4ff822e1d800 | lesswrong | bdHzoDJ3eA9MGJgmT | Burlington, VT - Spring ACX Meetup
Organized for ACX Spring Meetups 2023
Location: Battery Park, in the southern section of the park, near the William Wells statue. In the event of inclement weather we may relocate.
I will have an "ACX Meetup" sign.
Coordinates
Burlington LW/ACX Google Group
Comments
forrest-csu... | 172 |
stampy::4bab5c4a-893b-543b-bd5a-fd8964af8dcd | stampy | 49167ef8ffc46f188a1ab4bdccf7bd0f | DeepMind x UCL RL Lecture Series - Theoretical Fund. of Dynamic Programming Algorithms [4/13]
hi everybody and welcome back to our
fourth lecture on reinforcement learning
today we're going to go deeper into the
fundamentals of dynamic programming and
revisit some of the algorithms we've
introduced the last time we se... | 10,633 |
lesswrong::0c605ff0-796b-5c42-843a-cd424fe12d5d | lesswrong | 7eJ9Q6YxyB7LbiWSY | [SEQ RERUN] Sympathetic Minds
Today's post, Sympathetic Minds was originally published on 19 January 2009. A summary (taken from the LW wiki):
Mirror neurons are neurons that fire both when performing an action oneself, and watching someone else perform the same action - for example, a neuron that fires when you rais... | 825 |
lesswrong::d4665ced-4983-5df5-9c93-8658437fd2bb | lesswrong | ewkYgtZapQRtDPT2F | Additive Operations on Cartesian Frames
The mathematical object (but not the philosophical interpretation) of a Cartesian Frame is studied under the name "Chu space."
(In category theory, Chu spaces are usually studied in the special case of $W=2$. To learn more about Chu spaces, see Vaughan Pratt's guide to papers a... | 9,535 |
lesswrong::521ba4cf-5db2-5b7b-a07b-36f85aef8347 | lesswrong | 7HHKaD3BdsfHX7doz | 5 "Plan A" scenarios
https://ai-2040.com/ is a compelling look at one path forward in a world where the US decides to cooperate with China to slow down transformative AI, to buy us time to do it safely. (Ok, it has a bunch of potential paths forward, but one key path involving cooperation.)
To me, the key insight in ... | 1,633 |
End of preview. Expand in Data Studio
README.md exists but content is empty.
- Downloads last month
- 80