All 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 202320242025

All Jan Feb Mar Apr MayJunJul Aug Sep Oct Nov Dec

All 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 272829 30

Toward a taxonomy of cognitive benchmarks for agentic AGIs

Ben SmithJun 27, 2024, 11:50 PM

15 points

0 comments5 min readLW link

How Big a Deal are MatMul-Free Transformers?

JustisMillsJun 27, 2024, 10:28 PM

19 points

6 comments5 min readLW link

(justismills.substack.com)

Secondary forces of debt

KatjaGraceJun 27, 2024, 9:10 PM

81 points

18 comments2 min readLW link

(worldspiritsockpuppet.com)

Distillation of ‘Do language models plan for future tokens’

TheManxLoinerJun 27, 2024, 8:57 PM

26 points

2 comments6 min readLW link

how birds sense magnetic fields

bhauthJun 27, 2024, 6:59 PM

51 points

4 comments5 min readLW link

(www.bhauth.com)

Representation Tuning

Christopher AckermanJun 27, 2024, 5:44 PM

35 points

9 comments13 min readLW link

An issue with training schemers with supervised fine-tuning

Fabien RogerJun 27, 2024, 3:37 PM

49 points

12 comments6 min readLW link

AI #70: A Beautiful Sonnet

ZviJun 27, 2024, 2:40 PM

38 points

0 comments44 min readLW link

(thezvi.wordpress.com)

Detecting Genetically Engineered Viruses With Metagenomic Sequencing

jefftkJun 27, 2024, 2:01 PM

87 points

10 comments LW link

(naobservatory.org)

Cross Robin

jefftkJun 27, 2024, 3:10 AM

11 points

2 comments1 min readLW link

(www.jefftk.com)

Live Theory Part 0: Taking Intelligence Seriously

SahilJun 26, 2024, 9:37 PM

101 points

3 comments8 min readLW link

Instrumental vs Terminal Desiderata

Max HarmsJun 26, 2024, 8:57 PM

21 points

0 comments3 min readLW link

Imbue (Generally Intelligent) continue to make progress

Nathan Helm-BurgerJun 26, 2024, 8:41 PM

18 points

0 comments1 min readLW link

(imbue.com)

Tracing the steps

matimissonaJun 26, 2024, 7:22 PM

−8 points

2 comments4 min readLW link

Countering AI disinformation and deep fakes with digital signatures

Dave LindberghJun 26, 2024, 6:09 PM

13 points

5 comments1 min readLW link

Progress Conference 2024: Toward Abundant Futures

jasoncrawfordJun 26, 2024, 3:39 PM

40 points

2 comments1 min readLW link

(rootsofprogress.org)

Schelling points in the AGI policy space

mesaoptimizerJun 26, 2024, 1:19 PM

52 points

2 comments6 min readLW link

Bad lessons learned from the debate

bayesyatinaJun 26, 2024, 11:54 AM

8 points

5 comments6 min readLW link

Childhood and Education Roundup #6: College Edition

ZviJun 26, 2024, 11:40 AM

28 points

8 comments23 min readLW link

(thezvi.wordpress.com)

New fast transformer inference ASIC — Sohu by Etched

lemonhopeJun 26, 2024, 9:56 AM

8 points

9 comments1 min readLW link

(www.etched.com)

Empirical vs. Mathematical Joints of Nature

Elizabeth and Alex_Altair

Jun 26, 2024, 1:55 AM

35 points

1 comment5 min readLW link

My Current Claims and Cruxes on LLM Forecasting & Epistemics

ozziegooenJun 26, 2024, 12:40 AM

11 points

0 comments LW link

In favour of exploring nagging doubts about x-risk

owencbJun 25, 2024, 11:52 PM

105 points

2 comments LW link

What is a Tool?

johnswentworth and David Lorell

Jun 25, 2024, 11:40 PM

62 points

4 comments6 min readLW link

[Question] When do alignment researchers retire?

Jordan TaylorJun 25, 2024, 11:30 PM

4 points

2 comments1 min readLW link

Compute Governance Literature Review

sijarvisJun 25, 2024, 10:41 PM

11 points

0 comments13 min readLW link

Computational Complexity as an Intuition Pump for LLM Generality

aribrillJun 25, 2024, 8:25 PM

18 points

6 comments3 min readLW link

Failure Modes of Teaching AI Safety

Eleni AngelouJun 25, 2024, 7:07 PM

20 points

0 comments1 min readLW link

Kingfisher Summer Tour 2024

jefftkJun 25, 2024, 6:50 PM

9 points

0 comments1 min readLW link

(www.jefftk.com)

Incentive Learning vs Dead Sea Salt Experiment

Steven ByrnesJun 25, 2024, 5:49 PM

30 points

1 comment28 min readLW link

An Intuitive Explanation of Sparse Autoencoders for Mechanistic Interpretability of LLMs

Adam KarvonenJun 25, 2024, 3:57 PM

27 points

0 comments9 min readLW link

(adamkarvonen.github.io)

Formal verification, heuristic explanations and surprise accounting

Jacob_HiltonJun 25, 2024, 3:40 PM

156 points

11 comments9 min readLW link

(www.alignment.org)

Metastrategy get-started guide

TahpJun 25, 2024, 3:04 PM

6 points

1 comment8 min readLW link

Labor Participation is an Alignment Risk

alexJun 25, 2024, 2:15 PM

−5 points

2 comments17 min readLW link

Monthly Roundup #19: June 2024

ZviJun 25, 2024, 12:00 PM

28 points

9 comments54 min readLW link

(thezvi.wordpress.com)

Regularly meta-optimization

Crazy philosopherJun 25, 2024, 6:12 AM

−4 points

6 comments1 min readLW link

Memetics as an analogy and its implicit connotations

Rachel ShuJun 25, 2024, 5:13 AM

3 points

0 comments3 min readLW link

Mistakes people make when thinking about units

Isaac KingJun 25, 2024, 3:39 AM

74 points

14 comments7 min readLW link

Higher-effort summer solstice: What if we used AI (i.e., Angel Island)?

Rachel ShuJun 25, 2024, 1:35 AM

46 points

9 comments3 min readLW link

I’m a bit skeptical of AlphaFold 3

Oleg TrottJun 25, 2024, 12:04 AM

87 points

14 comments2 min readLW link

Being hella lost as rationality practice

Rachel ShuJun 24, 2024, 11:50 PM

14 points

0 comments2 min readLW link

A Basic Economics-Style Model of AI Existential Risk

Rubi J. HudsonJun 24, 2024, 8:26 PM

24 points

3 comments7 min readLW link

The Minority Coalition

Richard_NgoJun 24, 2024, 8:01 PM

103 points

9 comments5 min readLW link

(www.narrativeark.xyz)

Compact Proofs of Model Performance via Mechanistic Interpretability

LawrenceC, rajashree, Adrià Garriga-alonso and Jason Gross

24 Jun 2024 19:27 UTC

97 points

4 comments8 min readLW link

(arxiv.org)

Contrapositive Natural Abstraction—Project Intro

Elliot Callender24 Jun 2024 18:37 UTC

4 points

5 comments2 min readLW link

Sparse Features Through Time

Rogan Inglis24 Jun 2024 18:06 UTC

12 points

1 comment1 min readLW link

(roganinglis.io)

PSA: Consider alternatives to AUROC when reporting classifier metrics for alignment

rpglover6424 Jun 2024 17:53 UTC

18 points

1 comment3 min readLW link

Paying Russians to not invade Ukraine

djColliderBias24 Jun 2024 17:46 UTC

9 points

7 comments3 min readLW link

SAE feature geometry is outside the superposition hypothesis

jake_mendel24 Jun 2024 16:07 UTC

228 points

17 comments11 min readLW link

So you want to work on technical AI safety

gw24 Jun 2024 14:29 UTC

51 points

3 comments14 min readLW link