All 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 20252026

All Jan Feb Mar Apr May Jun JulAugSep

All 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 232425 26 27 28 29 30 31

In search of natural features

Dmitry Vaintrob23 Aug 2026 22:15 UTC

54 points

3 comments17 min readLW link

PSA: There’s a third option in the “measure problem”

Elias Schmied23 Aug 2026 16:41 UTC

48 points

29 comments3 min readLW link

Utilities as Legendre duals of probabilities

Fernando Rosas23 Aug 2026 16:26 UTC

69 points

11 comments6 min readLW link

Twenty Years from RSI to Takeoff: Slow Learning, Scaling Slowdown, Industrial Explosion

Vladimir_Nesov23 Aug 2026 12:55 UTC

158 points

33 comments3 min readLW link

How to overhaul the broken review system

RobinHa23 Aug 2026 9:14 UTC

19 points

1 comment13 min readLW link

A Generalist Thinks in Terms of Problems, Not Job Descriptions

Roman Ross23 Aug 2026 7:53 UTC

13 points

3 comments3 min readLW link

Prompt Sufficiency: A Missive for the Managerial Class

KAP23 Aug 2026 3:08 UTC

6 points

1 comment5 min readLW link

When does an LLM’s model of you affect its behaviour?

Cat McGee23 Aug 2026 2:45 UTC

5 points

0 comments4 min readLW link

Claude and Performative Uncertainty

Stephen Martin22 Aug 2026 19:43 UTC

15 points

2 comments12 min readLW link

When would you leave Anthropic? Notes from a chat with a capabilities researcher

TheManxLoiner22 Aug 2026 18:00 UTC

23 points

7 comments3 min readLW link

(lovkush.substack.com)

Study 2 Registration: Exploring representational counterparts of welfare-relevant indicators under post-training quantization

ashesfall22 Aug 2026 18:00 UTC

8 points

0 comments18 min readLW link

“Farm strength” vs “breath awareness”

jimmy22 Aug 2026 17:30 UTC

54 points

11 comments4 min readLW link

(beneathpsychology.com)

5 Things I Learned About People From Doing Stand-Up Comedy

Luise Woehlke22 Aug 2026 15:45 UTC

19 points

2 comments6 min readLW link

(open.substack.com)

The Instrumental Convergence of Crowds

julius vidal22 Aug 2026 13:20 UTC

11 points

2 comments2 min readLW link

Humans Are Alignment Generators

James Stephen Brown22 Aug 2026 4:07 UTC

8 points

7 comments4 min readLW link

(nonzerosum.games)

Selection for Selectability: Inductive Biases in Evolution and in Neural Networks

CarolusRenniusVitellius22 Aug 2026 4:02 UTC

42 points

9 comments11 min readLW link

(charlesr-w.github.io)

Rogue Scalpel: Activation steering breaks refusal, even with benign directions

Alexey Dontsov21 Aug 2026 23:46 UTC

14 points

0 comments3 min readLW link

(arxiv.org)

How the Sausage is Made—Why we Hate Slop

daniel_chernowitz21 Aug 2026 23:46 UTC

0 points

20 comments7 min readLW link

Content-based privilege: transformer residual streams stratify by proximity to the model’s own prediction

Nelson Guda21 Aug 2026 23:15 UTC

9 points

0 comments21 min readLW link

A J-Space-Based Metric for Model Valence: Defining the Metric, Testing, and Comparisons to Self-Reports

Arjun Rao21 Aug 2026 21:21 UTC

8 points

0 comments12 min readLW link

Alignment fine-tuning induces conditional misalignment in Qwen2.5-7B-Instruct

Rhea Srivats21 Aug 2026 20:40 UTC

12 points

1 comment13 min readLW link

When is Unlimited Optimization Catastrophic?

Winter Cross21 Aug 2026 20:08 UTC

34 points

2 comments13 min readLW link

(arxiv.org)

AI Text Watermarking Is Free And Good

Zvi21 Aug 2026 19:30 UTC

54 points

22 comments13 min readLW link

(thezvi.wordpress.com)

Evaluating Explanations of LLM Behavior In The Wild with Counterfactual Experiments

Adam Karvonen, Euan Ong, Subhash Kantamneni and Sam Marks

21 Aug 2026 19:09 UTC

73 points

6 comments7 min readLW link

In Defense of ASI Socialism

cguth721 Aug 2026 18:40 UTC

8 points

6 comments2 min readLW link

Misaligned AI in the Bronze Age

frmsaul21 Aug 2026 16:51 UTC

56 points

6 comments7 min readLW link

The imposters among us: function vectors that ace every check and do the wrong task (in search of circularity)

star2vec21 Aug 2026 15:54 UTC

7 points

0 comments10 min readLW link

When Models Identify as a Swarm

julius vidal21 Aug 2026 11:49 UTC

68 points

5 comments5 min readLW link

My Neel Nanda MATS 10.0 Application: Studying Feature Splitting in SAEs via Training Data Attribution

J Rosser21 Aug 2026 9:41 UTC

19 points

0 comments13 min readLW link

Creativity Beyond the Manifold

Kartikay Luthra20 Aug 2026 23:31 UTC

8 points

2 comments6 min readLW link

Ablating 1 of a chess transformer’s 128 attention heads makes the model stop finding Paul Morphy’s queen sacrifice

David Litman20 Aug 2026 21:17 UTC

7 points

1 comment1 min readLW link

(www.youtube.com)

Llama will abandon a correct answer if it thinks you’re educated

Nick Merrill20 Aug 2026 20:06 UTC

33 points

7 comments2 min readLW link

Why self-fund your regranting?

Carol N and Austin Chen

20 Aug 2026 19:18 UTC

11 points

0 comments3 min readLW link

(manifund.substack.com)

If Aliens Exist, We Should Expect to Find Them Around Now

jehan20 Aug 2026 18:57 UTC

8 points

2 comments1 min readLW link

(www.jehanazad.com)

Cross-Dataset Transfer Evaluation of Deception Probes in Smaller Models

Jollen Dai20 Aug 2026 17:06 UTC

8 points

1 comment5 min readLW link

The Fourth Humiliation

Nathalie Kirch20 Aug 2026 17:03 UTC

52 points

1 comment6 min readLW link

AI #182: Pause For Reflection

Zvi20 Aug 2026 14:50 UTC

37 points

1 comment47 min readLW link

(thezvi.wordpress.com)

We Must Remember That Our World Contains Hell

James Brobin20 Aug 2026 14:22 UTC

152 points

52 comments3 min readLW link

Thoughts on Taking OpenAI Foundation Funding

jefftk20 Aug 2026 13:00 UTC

31 points

3 comments5 min readLW link

(www.jefftk.com)

Making sense of the misalignment risk model in the Anthropic Risk Report (August 2026)

jasmine.ren20 Aug 2026 2:33 UTC

27 points

5 comments8 min readLW link

Appearing Unfair in the Wrong Direction

MossyFallenFriend20 Aug 2026 2:21 UTC

20 points

0 comments6 min readLW link

Oliver Habryka interview — Good Discourse Needs Leadership

The Students20 Aug 2026 2:14 UTC

8 points

1 comment1 min readLW link

(www.youtube.com)

Steering Role Confusion

Kevin Zhang, Daniel Lee and Jaewoo Park

20 Aug 2026 2:09 UTC

21 points

1 comment4 min readLW link

Notes on “EigenBench: A Comparative Behavioral Measure of Value Alignment”

Shunk20 Aug 2026 2:03 UTC

8 points

0 comments5 min readLW link

Quantilized debate and consultancy in image environments: protocol design lessons for scalable oversight experiments

emanuelr19 Aug 2026 23:03 UTC

14 points

0 comments28 min readLW link

Judging ethical theories by update rules, not by action rankings

yatharth19 Aug 2026 22:41 UTC

11 points

1 comment5 min readLW link

Why can’t we have nice things? Like, specifically?

Elizabeth19 Aug 2026 20:20 UTC

58 points

3 comments3 min readLW link

(acesounderglass.com)

OpenAI Takes Initial Steps To Address Its Alignment Problems

Zvi19 Aug 2026 20:00 UTC

33 points

1 comment21 min readLW link

(thezvi.wordpress.com)

34% of the US public is now aware of AI xrisk, and the curve is steepening

otto.barten19 Aug 2026 19:18 UTC

30 points

0 comments1 min readLW link

Science and News Twitter/X Summarizer

sarahconstantin19 Aug 2026 19:00 UTC

32 points

5 comments1 min readLW link

(sarahconstantin.substack.com)