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 MayJunJul Aug Sep Oct Nov Dec

All1 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 27 28 29 30

I would have shit in that alley, too

Declan Molony18 Jun 2024 4:41 UTC

436 points

134 comments4 min readLW link

Safety isn’t safety without a social model (or: dispelling the myth of per se technical safety)

Andrew_Critch14 Jun 2024 0:16 UTC

344 points

38 comments4 min readLW link

My AI Model Delta Compared To Yudkowsky

johnswentworth10 Jun 2024 16:12 UTC

277 points

102 comments4 min readLW link

Getting 50% (SoTA) on ARC-AGI with GPT-4o

ryan_greenblatt17 Jun 2024 18:44 UTC

262 points

50 comments13 min readLW link

SAE feature geometry is outside the superposition hypothesis

jake_mendel24 Jun 2024 16:07 UTC

221 points

17 comments11 min readLW link

LLM Generality is a Timeline Crux

eggsyntax24 Jun 2024 12:52 UTC

217 points

119 comments7 min readLW link

Response to Aschenbrenner’s “Situational Awareness”

Rob Bensinger6 Jun 2024 22:57 UTC

194 points

27 comments3 min readLW link

My AI Model Delta Compared To Christiano

johnswentworth12 Jun 2024 18:19 UTC

191 points

73 comments4 min readLW link

Two easy things that maybe Just Work to improve AI discourse

jacobjacob8 Jun 2024 15:51 UTC

189 points

35 comments2 min readLW link

Humming is not a free $100 bill

Elizabeth6 Jun 2024 20:10 UTC

183 points

6 comments3 min readLW link

(acesounderglass.com)

Boycott OpenAI

PeterMcCluskey18 Jun 2024 19:52 UTC

163 points

26 comments1 min readLW link

(bayesianinvestor.com)

Announcing ILIAD — Theoretical AI Alignment Conference

Nora_Ammann and Alexander Gietelink Oldenziel

5 Jun 2024 9:37 UTC

162 points

18 comments2 min readLW link

Sycophancy to subterfuge: Investigating reward tampering in large language models

Carson Denison and evhub

17 Jun 2024 18:41 UTC

161 points

22 comments8 min readLW link

(arxiv.org)

Connecting the Dots: LLMs can Infer & Verbalize Latent Structure from Training Data

Johannes Treutlein and Owain_Evans

21 Jun 2024 15:54 UTC

160 points

13 comments8 min readLW link

(arxiv.org)

Formal verification, heuristic explanations and surprise accounting

Jacob_Hilton25 Jun 2024 15:40 UTC

156 points

11 comments9 min readLW link

(www.alignment.org)

The Incredible Fentanyl-Detecting Machine

sarahconstantin28 Jun 2024 22:10 UTC

154 points

26 comments7 min readLW link

(sarahconstantin.substack.com)

0. CAST: Corrigibility as Singular Target

Max Harms7 Jun 2024 22:29 UTC

144 points

12 comments8 min readLW link

Loving a world you don’t trust

Joe Carlsmith18 Jun 2024 19:31 UTC

134 points

13 comments33 min readLW link

How it All Went Down: The Puzzle Hunt that took us way, way Less Online

A*2 Jun 2024 8:01 UTC

134 points

5 comments5 min readLW link

Why I don’t believe in the placebo effect

transhumanist_atom_understander10 Jun 2024 2:37 UTC

131 points

22 comments9 min readLW link

[Question] What do coherence arguments actually prove about agentic behavior?

sunwillrise1 Jun 2024 9:37 UTC

123 points

35 comments6 min readLW link

Evidence of Learned Look-Ahead in a Chess-Playing Neural Network

Erik Jenner4 Jun 2024 15:50 UTC

120 points

14 comments13 min readLW link

AI catastrophes and rogue deployments

Buck3 Jun 2024 17:04 UTC

119 points

16 comments8 min readLW link

The Standard Analogy

Zack_M_Davis3 Jun 2024 17:15 UTC

118 points

28 comments12 min readLW link

Anthropic’s Certificate of Incorporation

Zach Stein-Perlman12 Jun 2024 13:00 UTC

115 points

4 comments4 min readLW link

The Leopold Model: Analysis and Reactions

Zvi14 Jun 2024 15:10 UTC

108 points

19 comments57 min readLW link

(thezvi.wordpress.com)

Demystifying “Alignment” through a Comic

milanrosko9 Jun 2024 8:24 UTC

106 points

19 comments1 min readLW link

Scaling and evaluating sparse autoencoders

leogao6 Jun 2024 22:50 UTC

106 points

6 comments1 min readLW link

In favour of exploring nagging doubts about x-risk

owencb25 Jun 2024 23:52 UTC

105 points

2 comments1 min readLW link

Live Theory Part 0: Taking Intelligence Seriously

Sahil26 Jun 2024 21:37 UTC

101 points

3 comments8 min readLW link

On Dwarksh’s Podcast with Leopold Aschenbrenner

Zvi10 Jun 2024 12:40 UTC

101 points

7 comments59 min readLW link

(thezvi.wordpress.com)

The Minority Coalition

Richard_Ngo24 Jun 2024 20:01 UTC

99 points

7 comments5 min readLW link

(www.narrativeark.xyz)

CIV: a story

Richard_Ngo15 Jun 2024 22:36 UTC

98 points

6 comments9 min readLW link

(www.narrativeark.xyz)

OpenAI #8: The Right to Warn

Zvi17 Jun 2024 12:00 UTC

97 points

8 comments34 min readLW link

(thezvi.wordpress.com)

Access to powerful AI might make computer security radically easier

Buck8 Jun 2024 6:00 UTC

97 points

14 comments6 min readLW link

Comments on Anthropic’s Scaling Monosemanticity

Robert_AIZI3 Jun 2024 12:15 UTC

97 points

8 comments7 min readLW link

Compact Proofs of Model Performance via Mechanistic Interpretability

LawrenceC, rajashree, Adrià Garriga-alonso and Jason Gross

24 Jun 2024 19:27 UTC

96 points

3 comments8 min readLW link

(arxiv.org)

On Claude 3.5 Sonnet

Zvi24 Jun 2024 12:00 UTC

95 points

14 comments13 min readLW link

(thezvi.wordpress.com)

Ilya Sutskever created a new AGI startup

harfe19 Jun 2024 17:17 UTC

95 points

35 comments1 min readLW link

(ssi.inc)

Towards a Less Bullshit Model of Semantics

johnswentworth and David Lorell

17 Jun 2024 15:51 UTC

94 points

44 comments21 min readLW link

Takeoff speeds presentation at Anthropic

Tom Davidson4 Jun 2024 22:46 UTC

92 points

0 comments25 min readLW link

Just admit that you’ve zoned out

joec4 Jun 2024 2:51 UTC

91 points

22 comments2 min readLW link

Quotes from Leopold Aschenbrenner’s Situational Awareness Paper

Zvi7 Jun 2024 11:40 UTC

91 points

10 comments37 min readLW link

(thezvi.wordpress.com)

I’m a bit skeptical of AlphaFold 3

Oleg Trott25 Jun 2024 0:04 UTC

87 points

14 comments2 min readLW link

Detecting Genetically Engineered Viruses With Metagenomic Sequencing

jefftk27 Jun 2024 14:01 UTC

87 points

10 comments1 min readLW link

(naobservatory.org)

[Paper] Stress-testing capability elicitation with password-locked models

Fabien Roger and ryan_greenblatt

4 Jun 2024 14:52 UTC

85 points

10 comments12 min readLW link

(arxiv.org)

[Paper] AI Sandbagging: Language Models can Strategically Underperform on Evaluations

Teun van der Weij, Felix Hofstätter, Ollie J, Sam F. Brown and Francis Rhys Ward

13 Jun 2024 10:04 UTC

84 points

10 comments2 min readLW link

(arxiv.org)

Actually, Power Plants May Be an AI Training Bottleneck.

Lao Mein20 Jun 2024 4:41 UTC

83 points

13 comments2 min readLW link

Corrigibility = Tool-ness?

johnswentworth and David Lorell

28 Jun 2024 1:19 UTC

78 points

8 comments9 min readLW link

AI takeoff and nuclear war

owencb11 Jun 2024 19:36 UTC

78 points

6 comments11 min readLW link

(strangecities.substack.com)