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 Sep Oct NovDec

All 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 181920 21 22 23 24 25 26 27 28 29 30 31

A Solution for AGI/ASI Safety

Weibing WangDec 18, 2024, 7:44 PM

50 points

25 votes

Overall karma indicates overall quality.

29 comments1 min readLW link

Takes on “Alignment Faking in Large Language Models”

Joe CarlsmithDec 18, 2024, 6:22 PM

105 points

30 votes

Overall karma indicates overall quality.

7 comments62 min readLW link

A Matter of Taste

ZviDec 18, 2024, 5:50 PM

36 points

12 votes

Overall karma indicates overall quality.

4 comments11 min readLW link

(thezvi.wordpress.com)

Are we a different person each time? A simple argument for the impermanence of our identity

l4mpDec 18, 2024, 5:21 PM

−4 points

5 votes

Overall karma indicates overall quality.

5 comments1 min readLW link

Alignment Faking in Large Language Models

ryan_greenblatt, evhub, Carson Denison, Benjamin Wright, Fabien Roger, Monte M, Sam Marks, Johannes Treutlein, Sam Bowman and Buck

Dec 18, 2024, 5:19 PM

489 points

177 votes

Overall karma indicates overall quality.

75 comments10 min readLW link

Can o1-preview find major mistakes amongst 59 NeurIPS ’24 MLSB papers?

Abhishaike MahajanDec 18, 2024, 2:21 PM

19 points

11 votes

Overall karma indicates overall quality.

0 comments6 min readLW link

(www.owlposting.com)

Walking Sue

Matthew McRedmondDec 18, 2024, 1:19 PM

2 points

4 votes

Overall karma indicates overall quality.

5 comments8 min readLW link

What conclusions can be drawn from a single observation about wealth in tennis?

Trevor CappalloDec 18, 2024, 9:55 AM

8 points

5 votes

Overall karma indicates overall quality.

3 comments2 min readLW link

Don’t Associate AI Safety With Activism

EneaszDec 18, 2024, 8:01 AM

17 points

25 votes

Overall karma indicates overall quality.

15 comments1 min readLW link

(deathisbad.substack.com)

[Question] How should I optimize my decision making model for ‘ideas’?

CstineSublimeDec 18, 2024, 4:09 AM

3 points

1 vote

Overall karma indicates overall quality.

0 comments4 min readLW link

Preppers Are Too Negative on Objects

jefftkDec 18, 2024, 2:30 AM

44 points

17 votes

Overall karma indicates overall quality.

2 comments1 min readLW link

(www.jefftk.com)

Review: Breaking Free with Dr. Stone

TurnTroutDec 18, 2024, 1:26 AM

47 points

23 votes

Overall karma indicates overall quality.

5 comments1 min readLW link

(turntrout.com)

Ablations for “Frontier Models are Capable of In-context Scheming”

AlexMeinke, Bronson Schoen, Marius Hobbhahn, Mikita Balesni, Jérémy Scheurer and rusheb

Dec 17, 2024, 11:58 PM

115 points

35 votes

Overall karma indicates overall quality.

1 comment2 min readLW link

Careless thinking: A theory of bad thinking

Nathan YoungDec 17, 2024, 6:23 PM

49 points

16 votes

Overall karma indicates overall quality.

17 comments9 min readLW link

(nathanpmyoung.substack.com)

The Second Gemini

ZviDec 17, 2024, 3:50 PM

23 points

9 votes

Overall karma indicates overall quality.

0 comments11 min readLW link

(thezvi.wordpress.com)

AIS Hungary is hiring a part-time Technical Lead! (Deadline: Dec 31st)

gergogasparDec 17, 2024, 2:12 PM

1 point

1 vote

Overall karma indicates overall quality.

0 comments2 min readLW link

Everything you care about is in the map

TahpDec 17, 2024, 2:05 PM

17 points

12 votes

Overall karma indicates overall quality.

27 comments3 min readLW link

Reality is Fractal-Shaped

silentbobDec 17, 2024, 1:52 PM

18 points

12 votes

Overall karma indicates overall quality.

1 comment8 min readLW link

Trying to translate when people talk past each other

Kaj_SotalaDec 17, 2024, 9:40 AM

41 points

17 votes

Overall karma indicates overall quality.

12 comments6 min readLW link

(kajsotala.fi)

What is “wireheading”?

Vishakha and Algon

Dec 17, 2024, 7:49 AM

10 points

3 votes

Overall karma indicates overall quality.

0 comments1 min readLW link

(aisafety.info)

1 What If We Rebuild Motivation with the Fermi ESTIMATion?

P. JoãoDec 17, 2024, 7:46 AM

6 points

6 votes

Overall karma indicates overall quality.

0 comments3 min readLW link

Where do you put your ideas?

CstineSublimeDec 17, 2024, 7:26 AM

9 points

5 votes

Overall karma indicates overall quality.

20 comments1 min readLW link

Elevating Air Purifiers

jefftkDec 17, 2024, 1:40 AM

25 points

5 votes

Overall karma indicates overall quality.

0 comments1 min readLW link

(www.jefftk.com)

A dataset of questions on decision-theoretic reasoning in Newcomb-like problems

Caspar Oesterheld, Ethan Perez and Chi Nguyen

Dec 16, 2024, 10:42 PM

50 points

27 votes

Overall karma indicates overall quality.

1 comment2 min readLW link

(arxiv.org)

A practical guide to tiling the universe with hedonium

Vittu PerkeleDec 16, 2024, 9:25 PM

−8 points

4 votes

Overall karma indicates overall quality.

1 comment1 min readLW link

(perkeleperusing.substack.com)

AI Safety Seed Funding Network—Join as a Donor or Investor

Alexandra BosDec 16, 2024, 7:30 PM

30 points

9 votes

Overall karma indicates overall quality.

0 comments2 min readLW link

I read every major AI lab’s safety plan so you don’t have to

sarahhwDec 16, 2024, 6:51 PM

20 points

5 votes

Overall karma indicates overall quality.

0 comments12 min readLW link

(longerramblings.substack.com)

Grokking revisited: reverse engineering grokking modulo addition in LSTM

Nikita Khomich and Danik

Dec 16, 2024, 6:48 PM

4 points

4 votes

Overall karma indicates overall quality.

0 comments6 min readLW link

Progress links and short notes, 2024-12-16

jasoncrawfordDec 16, 2024, 5:24 PM

7 points

2 votes

Overall karma indicates overall quality.

0 comments2 min readLW link

(newsletter.rootsofprogress.org)

Effective Altruism FAQ

Bentham's BulldogDec 16, 2024, 4:27 PM

0 points

9 votes

Overall karma indicates overall quality.

7 comments12 min readLW link

Variably compressibly studies are fun

dkl9Dec 16, 2024, 4:00 PM

0 points

5 votes

Overall karma indicates overall quality.

0 comments2 min readLW link

(dkl9.net)

AIs Will Increasingly Attempt Shenanigans

ZviDec 16, 2024, 3:20 PM

118 points

59 votes

Overall karma indicates overall quality.

2 comments26 min readLW link

(thezvi.wordpress.com)

Testing which LLM architectures can do hidden serial reasoning

Filip SondejDec 16, 2024, 1:48 PM

84 points

29 votes

Overall karma indicates overall quality.

9 comments4 min readLW link

NeuroAI for AI safety: A Differential Path

nz and Patrick Mineault

Dec 16, 2024, 1:17 PM

22 points

6 votes

Overall karma indicates overall quality.

0 comments7 min readLW link

(arxiv.org)

Circling as practice for “just be yourself”

Kaj_SotalaDec 16, 2024, 7:40 AM

87 points

40 votes

Overall karma indicates overall quality.

6 comments4 min readLW link

(kajsotala.fi)

Reanalyzing the 2023 Expert Survey on Progress in AI

AI ImpactsDec 16, 2024, 6:10 AM

8 points

2 votes

Overall karma indicates overall quality.

0 comments1 min readLW link

(blog.aiimpacts.org)

Ideas for benchmarking LLM creativity

gwernDec 16, 2024, 5:18 AM

60 points

23 votes

Overall karma indicates overall quality.

11 comments1 min readLW link

(gwern.net)

Comparing the AirFanta 3Pro to the Coway AP-1512

jefftkDec 16, 2024, 1:40 AM

13 points

3 votes

Overall karma indicates overall quality.

0 comments1 min readLW link

(www.jefftk.com)

[Question] are IQ tests a good measure of intelligence?

KvmanThinkingDec 15, 2024, 11:06 PM

0 points

8 votes

Overall karma indicates overall quality.

5 comments1 min readLW link

Madison Secular Solstice

svfritzDec 15, 2024, 9:52 PM

1 point

1 vote

Overall karma indicates overall quality.

0 comments1 min readLW link

[Question] Is AI alignment a purely functional property?

RokoDec 15, 2024, 9:42 PM

13 points

5 votes

Overall karma indicates overall quality.

8 comments1 min readLW link

[Question] How counterfactual are logical counterfactuals?

Donald HobsonDec 15, 2024, 9:16 PM

11 points

7 votes

Overall karma indicates overall quality.

10 comments1 min readLW link

Debunking the myth of safe AI

henophiliaDec 15, 2024, 5:44 PM

−11 points

5 votes

Overall karma indicates overall quality.

8 comments1 min readLW link

(henophilia.substack.com)

Introducing Avatarism: A Rational Framework for Building actual Heaven

ratiba roDec 15, 2024, 5:17 PM

2 points

4 votes

Overall karma indicates overall quality.

2 comments2 min readLW link

A Public Choice Take on Effective Altruism

vaishnav92Dec 15, 2024, 4:58 PM

9 points

3 votes

Overall karma indicates overall quality.

4 comments3 min readLW link

(www.optimaloutliers.com)

World Models I’m Currently Building

temporaryDec 15, 2024, 4:29 PM

5 points

4 votes

Overall karma indicates overall quality.

1 comment1 min readLW link

(samuelshadrach.com)

Dress Up For Secular Solstice

Gordon H.S.Dec 15, 2024, 4:28 PM

33 points

21 votes

Overall karma indicates overall quality.

13 comments7 min readLW link

Remap your caps lock key

bilalchughtaiDec 15, 2024, 2:03 PM

81 points

59 votes

Overall karma indicates overall quality.

21 comments1 min readLW link

Effective Evil’s AI Misalignment Plan

lsusrDec 15, 2024, 7:39 AM

83 points

57 votes

Overall karma indicates overall quality.

9 comments3 min readLW link

How to Edit an Essay into a Solstice Speech?

CzynskiDec 15, 2024, 4:30 AM

5 points

2 votes

Overall karma indicates overall quality.

1 comment1 min readLW link

(thepdv.wordpress.com)