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 May Jun Jul Aug SepOctNov Dec

All 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 222324 25 26 27 28 29 30 31

Monosemanticity & Quantization

Rahul ChandOct 22, 2024, 10:57 PM

1 point

3 votes

Overall karma indicates overall quality.

0 comments9 min readLW link

[Question] What is the alpha in one bit of evidence?

J BostockOct 22, 2024, 9:57 PM

20 points

6 votes

Overall karma indicates overall quality.

13 comments1 min readLW link

Catastrophic sabotage as a major threat model for human-level AI systems

evhubOct 22, 2024, 8:57 PM

93 points

44 votes

Overall karma indicates overall quality.

13 comments15 min readLW link

Why I quit effective altruism, and why Timothy Telleen-Lawton is staying (for now)

ElizabethOct 22, 2024, 6:20 PM

76 points

42 votes

Overall karma indicates overall quality.

82 comments1 min readLW link

(acesounderglass.com)

Decision-Making Under Uncertainty: Lessons From AI

JonasbOct 22, 2024, 5:54 PM

−1 points

2 votes

Overall karma indicates overall quality.

0 comments5 min readLW link

(www.denominations.io)

Testing Genetic Engineering Detection with Spike-Ins

jefftkOct 22, 2024, 5:20 PM

9 points

1 vote

Overall karma indicates overall quality.

0 comments4 min readLW link

(naobservatory.org)

Predictions as Public Works Project — What Metaculus Is Building Next

ChristianWilliamsOct 22, 2024, 4:35 PM

5 points

2 votes

Overall karma indicates overall quality.

0 comments5 min readLW link

(www.metaculus.com)

Gorges of gender on a terrain of traits

dkl9Oct 22, 2024, 4:18 PM

−7 points

6 votes

Overall karma indicates overall quality.

1 comment3 min readLW link

(dkl9.net)

A Defense of Peer Review

Niko_McCarty and delton137

Oct 22, 2024, 4:16 PM

23 points

12 votes

Overall karma indicates overall quality.

1 comment22 min readLW link

(www.asimov.press)

BIG-Bench Canary Contamination in GPT-4

JozdienOct 22, 2024, 3:40 PM

130 points

66 votes

Overall karma indicates overall quality.

18 comments4 min readLW link

[Paper Blogpost] When Your AIs Deceive You: Challenges with Partial Observability in RLHF

Leon LangOct 22, 2024, 1:57 PM

51 points

22 votes

Overall karma indicates overall quality.

2 comments18 min readLW link

(arxiv.org)

[Intuitive self-models] 6. Awakening / Enlightenment / PNSE

Steven ByrnesOct 22, 2024, 1:23 PM

73 points

18 votes

Overall karma indicates overall quality.

11 comments21 min readLW link

Resolving von Neumann-Morgenstern Inconsistent Preferences

niplavOct 22, 2024, 11:45 AM

39 points

13 votes

Overall karma indicates overall quality.

5 comments58 min readLW link

Lenses of Control

WillPetilloOct 22, 2024, 7:51 AM

14 points

7 votes

Overall karma indicates overall quality.

0 comments9 min readLW link

A Brief Explanation of AI Control

Aaron_ScherOct 22, 2024, 7:00 AM

8 points

4 votes

Overall karma indicates overall quality.

1 comment6 min readLW link

Longevity, AI, and Cognitive Research Hackathon @ MIT

ekkoláptoOct 22, 2024, 6:19 AM

1 point

1 vote

Overall karma indicates overall quality.

0 comments1 min readLW link

Conversational Signposts—How to stop having boring social interactions

Declan MolonyOct 22, 2024, 5:37 AM

13 points

6 votes

Overall karma indicates overall quality.

6 comments2 min readLW link

I got dysentery so you don’t have to

eukaryoteOct 22, 2024, 4:55 AM

324 points

163 votes

Overall karma indicates overall quality.

6 comments17 min readLW link

(eukaryotewritesblog.com)

Transformers Explained (Again)

RohanSOct 22, 2024, 4:06 AM

4 points

2 votes

Overall karma indicates overall quality.

0 comments18 min readLW link

Sleeping on Stage

jefftkOct 22, 2024, 12:50 AM

26 points

9 votes

Overall karma indicates overall quality.

3 comments1 min readLW link

(www.jefftk.com)

The Mask Comes Off: At What Price?

ZviOct 21, 2024, 11:50 PM

72 points

28 votes

Overall karma indicates overall quality.

16 comments8 min readLW link

(thezvi.wordpress.com)

Distinguishing ways AI can be “concentrated”

Matthew BarnettOct 21, 2024, 10:21 PM

34 points

12 votes

Overall karma indicates overall quality.

2 comments4 min readLW link

Jailbreaking ChatGPT and Claude using Web API Context Injection

Jaehyuk LimOct 21, 2024, 9:34 PM

4 points

7 votes

Overall karma indicates overall quality.

0 comments3 min readLW link

How to Teach Your Brain to Hate Procrastination

10xyzOct 21, 2024, 8:12 PM

3 points

4 votes

Overall karma indicates overall quality.

0 comments2 min readLW link

Pausing for what?

MountainPathOct 21, 2024, 8:12 PM

0 points

4 votes

Overall karma indicates overall quality.

1 comment1 min readLW link

What is autonomy? Why boundaries are necessary.

Chris LakinOct 21, 2024, 5:56 PM

8 points

6 votes

Overall karma indicates overall quality.

1 comment1 min readLW link

(chrislakin.blog)

Could randomly choosing people to serve as representatives lead to better government?

John HuangOct 21, 2024, 5:10 PM

76 points

42 votes

Overall karma indicates overall quality.

13 comments10 min readLW link

There aren’t enough smart people in biology doing something boring

Abhishaike MahajanOct 21, 2024, 3:52 PM

28 points

15 votes

Overall karma indicates overall quality.

13 comments10 min readLW link

Automation collapse

Geoffrey Irving, Tomek Korbak and Benjamin Hilton

Oct 21, 2024, 2:50 PM

72 points

25 votes

Overall karma indicates overall quality.

9 comments7 min readLW link

What AI companies should do: Some rough ideas

Zach Stein-PerlmanOct 21, 2024, 2:00 PM

33 points

11 votes

Overall karma indicates overall quality.

10 comments5 min readLW link

[Question] What should OpenAI do that it hasn’t already done, to stop their vacancies from being advertised on the 80k Job Board?

WitheringWeightsOct 21, 2024, 1:57 PM

22 points

8 votes

Overall karma indicates overall quality.

0 comments1 min readLW link

A Rocket–Interpretability Analogy

plexOct 21, 2024, 1:55 PM

156 points

64 votes

Overall karma indicates overall quality.

31 comments1 min readLW link

Tokyo AI Safety 2025: Call For Papers

BlaineOct 21, 2024, 8:43 AM

24 points

5 votes

Overall karma indicates overall quality.

0 comments3 min readLW link

(www.tais2025.cc)

OpenAI defected, but we can take honest actions

RemmeltOct 21, 2024, 8:41 AM

17 points

16 votes

Overall karma indicates overall quality.

16 comments2 min readLW link

Slightly More Than You Wanted To Know: Pregnancy Length Effects

JustisMillsOct 21, 2024, 1:26 AM

63 points

22 votes

Overall karma indicates overall quality.

4 comments5 min readLW link

(justismills.substack.com)

Information vs Assurance

johnswentworthOct 20, 2024, 11:16 PM

187 points

102 votes

Overall karma indicates overall quality.

18 comments2 min readLW link

Liquid vs Illiquid Careers

vaishnav92Oct 20, 2024, 11:03 PM

35 points

24 votes

Overall karma indicates overall quality.

7 comments7 min readLW link

(vaishnavsunil.substack.com)

AI Can be “Gradient Aware” Without Doing Gradient hacking.

SodiumOct 20, 2024, 9:02 PM

21 points

10 votes

Overall karma indicates overall quality.

1 comment2 min readLW link

A brief theory of why we think things are good or bad

David JohnstonOct 20, 2024, 8:31 PM

7 points

5 votes

Overall karma indicates overall quality.

10 comments4 min readLW link

Thinking in 2D

sarahconstantinOct 20, 2024, 7:30 PM

27 points

7 votes

Overall karma indicates overall quality.

0 comments8 min readLW link

(sarahconstantin.substack.com)

Podcast discussing Hanson’s Cultural Drift Argument

vaishnav92 and regan.arntz.gray

Oct 20, 2024, 5:58 PM

3 points

3 votes

Overall karma indicates overall quality.

0 comments1 min readLW link

(moralmayhem.substack.com)

Advice on Communicating Concisely

EvolutionByDesignOct 20, 2024, 4:45 PM

3 points

3 votes

Overall karma indicates overall quality.

9 comments1 min readLW link

Ambiguities or the issues we face with AI in medicine

Thehumanproject.aiOct 20, 2024, 4:45 PM

2 points

2 votes

Overall karma indicates overall quality.

0 comments5 min readLW link

The Personal Implications of AGI Realism

xiznebOct 20, 2024, 4:43 PM

7 points

9 votes

Overall karma indicates overall quality.

8 comments5 min readLW link

Safety tax functions

owencbOct 20, 2024, 2:08 PM

31 points

8 votes

Overall karma indicates overall quality.

0 comments6 min readLW link

(strangecities.substack.com)

Exploring the Platonic Representation Hypothesis Beyond In-Distribution Data

rokosbasiliskOct 20, 2024, 8:40 AM

12 points

5 votes

Overall karma indicates overall quality.

2 comments1 min readLW link

Electoral Systems

RedFishBlueFishOct 20, 2024, 3:25 AM

1 point

3 votes

Overall karma indicates overall quality.

0 comments14 min readLW link

Overcoming Bias Anthology

Arjun PanicksseryOct 20, 2024, 2:01 AM

169 points

61 votes

Overall karma indicates overall quality.

14 comments2 min readLW link

(overcoming-bias-anthology.com)

D/acc AI Security Salon

Allison DuettmannOct 19, 2024, 10:17 PM

19 points

6 votes

Overall karma indicates overall quality.

0 comments1 min readLW link

Who Should Have Been Killed, and Contains Neato? Who Else Could It Be, but that Villain Magneto!

Ace DelgadoOct 19, 2024, 8:39 PM

−16 points

8 votes

Overall karma indicates overall quality.

0 comments1 min readLW link