All 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 20232024

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

A Solution for AGI/ASI Safety

Weibing Wang18 Dec 2024 19:44 UTC

41 points

10 comments1 min readLW link

Takes on “Alignment Faking in Large Language Models”

Joe Carlsmith18 Dec 2024 18:22 UTC

92 points

8 comments62 min readLW link

A Matter of Taste

Zvi18 Dec 2024 17:50 UTC

36 points

4 comments11 min readLW link

(thezvi.wordpress.com)

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

l4mp18 Dec 2024 17:21 UTC

−4 points

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

18 Dec 2024 17:19 UTC

413 points

51 comments10 min readLW link

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

Abhishaike Mahajan18 Dec 2024 14:21 UTC

18 points

0 comments6 min readLW link

(www.owlposting.com)

Walking Sue

Matthew McRedmond18 Dec 2024 13:19 UTC

2 points

5 comments8 min readLW link

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

Trevor Cappallo18 Dec 2024 9:55 UTC

8 points

3 comments2 min readLW link

Don’t Associate AI Safety With Activism

Eneasz18 Dec 2024 8:01 UTC

16 points

13 comments1 min readLW link

(deathisbad.substack.com)

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

CstineSublime18 Dec 2024 4:09 UTC

3 points

0 comments4 min readLW link

Preppers Are Too Negative on Objects

jefftk18 Dec 2024 2:30 UTC

40 points

2 comments1 min readLW link

(www.jefftk.com)

Review: Breaking Free with Dr. Stone

TurnTrout18 Dec 2024 1:26 UTC

49 points

3 comments1 min readLW link

(turntrout.com)

Being Present is Not a Skill

Chipmonk18 Dec 2024 1:11 UTC

17 points

6 comments1 min readLW link

(chrislakin.blog)

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

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

17 Dec 2024 23:58 UTC

107 points

1 comment2 min readLW link

bending light

Recurrented17 Dec 2024 22:40 UTC

1 point

5 comments3 min readLW link

(futuring.substack.com)

Careless thinking: A theory of bad thinking

Nathan Young17 Dec 2024 18:23 UTC

43 points

17 comments9 min readLW link

(nathanpmyoung.substack.com)

The Second Gemini

Zvi17 Dec 2024 15:50 UTC

23 points

0 comments11 min readLW link

(thezvi.wordpress.com)

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

gergogaspar17 Dec 2024 14:12 UTC

1 point

0 comments2 min readLW link

Everything you care about is in the map

Tahp17 Dec 2024 14:05 UTC

17 points

26 comments3 min readLW link

Reality is Fractal-Shaped

silentbob17 Dec 2024 13:52 UTC

17 points

1 comment8 min readLW link

Trying to translate when people talk past each other

Kaj_Sotala17 Dec 2024 9:40 UTC

41 points

12 comments6 min readLW link

(kajsotala.fi)

What is “wireheading”?

Vishakha17 Dec 2024 7:49 UTC

10 points

0 comments1 min readLW link

(aisafety.info)

3 What If We Could Map Our Motivation as Channels of Flow?

P. João17 Dec 2024 7:47 UTC

3 points

0 comments6 min readLW link

2 What if Life Comes with a Natural Calibration to Estimate you?

P. João17 Dec 2024 7:47 UTC

1 point

0 comments10 min readLW link

1 What If We Rebuild Motivation with the Fermi ESTIMATion?

P. João17 Dec 2024 7:46 UTC

5 points

0 comments3 min readLW link

Where do you put your ideas?

CstineSublime17 Dec 2024 7:26 UTC

7 points

12 comments1 min readLW link

Elevating Air Purifiers

jefftk17 Dec 2024 1:40 UTC

25 points

0 comments1 min readLW link

(www.jefftk.com)

0 Motivation Mapping through Information Theory

P. João16 Dec 2024 23:17 UTC

8 points

0 comments28 min readLW link

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

Caspar Oesterheld, Ethan Perez and Chi Nguyen

16 Dec 2024 22:42 UTC

45 points

1 comment2 min readLW link

(arxiv.org)

A practical guide to tiling the universe with hedonium

Vittu Perkele16 Dec 2024 21:25 UTC

−9 points

1 comment1 min readLW link

(perkeleperusing.substack.com)

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

Alexandra Bos16 Dec 2024 19:30 UTC

30 points

0 comments1 min readLW link

Is this a better way to do matchmaking?

Chipmonk16 Dec 2024 19:06 UTC

9 points

4 comments1 min readLW link

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

sarahhw16 Dec 2024 18:51 UTC

20 points

0 comments12 min readLW link

(longerramblings.substack.com)

Grokking revisited: reverse engineering grokking modulo addition in LSTM

Nikita Khomich and Danik

16 Dec 2024 18:48 UTC

4 points

0 comments6 min readLW link

[Question] Do infinite alternatives make AI alignment impossible?

Dakara16 Dec 2024 18:11 UTC

11 points

2 comments1 min readLW link

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

jasoncrawford16 Dec 2024 17:24 UTC

7 points

0 comments2 min readLW link

(newsletter.rootsofprogress.org)

Effective Altruism FAQ

omnizoid16 Dec 2024 16:27 UTC

0 points

6 comments12 min readLW link

Variably compressibly studies are fun

dkl916 Dec 2024 16:00 UTC

0 points

0 comments2 min readLW link

(dkl9.net)

AIs Will Increasingly Attempt Shenanigans

Zvi16 Dec 2024 15:20 UTC

94 points

4 comments26 min readLW link

(thezvi.wordpress.com)

Testing which LLM architectures can do hidden serial reasoning

Filip Sondej16 Dec 2024 13:48 UTC

79 points

9 comments4 min readLW link

NeuroAI for AI safety: A Differential Path

nz and Patrick Mineault

16 Dec 2024 13:17 UTC

14 points

0 comments7 min readLW link

(arxiv.org)

Circling as practice for “just be yourself”

Kaj_Sotala16 Dec 2024 7:40 UTC

83 points

4 comments4 min readLW link

(kajsotala.fi)

Reanalyzing the 2023 Expert Survey on Progress in AI

AI Impacts16 Dec 2024 6:10 UTC

8 points

0 comments1 min readLW link

(blog.aiimpacts.org)

Ideas for benchmarking LLM creativity

gwern16 Dec 2024 5:18 UTC

50 points

10 comments1 min readLW link

(gwern.net)

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

jefftk16 Dec 2024 1:40 UTC

13 points

0 comments1 min readLW link

(www.jefftk.com)

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

KvmanThinking15 Dec 2024 23:06 UTC

0 points

5 comments1 min readLW link

Madison Secular Solstice

svfritz15 Dec 2024 21:52 UTC

1 point

0 comments1 min readLW link

[Question] Is AI alignment a purely functional property?

Roko15 Dec 2024 21:42 UTC

13 points

7 comments1 min readLW link

[Question] How counterfactual are logical counterfactuals?

Donald Hobson15 Dec 2024 21:16 UTC

11 points

10 comments1 min readLW link

Debunking the myth of safe AI

henophilia15 Dec 2024 17:44 UTC

−11 points

7 comments1 min readLW link

(henophilia.substack.com)