Refusal in LLMs is mediated by a single direction

Andy Arditi, Oscar Obeso, Aaquib111, wesg and Neel Nanda

27 Apr 2024 11:13 UTC

185 points

79 comments10 min readLW link

Key takeaways from our EA and alignment research surveys

Cameron Berg, Judd Rosenblatt, florin_pop and AE Studio

3 May 2024 18:10 UTC

87 points

9 comments21 min readLW link

Can we build a better Public Doublecrux?

Raemon11 May 2024 19:21 UTC

10 points

0 comments4 min readLW link

MATS Winter 2023-24 Retrospective

Rocket, Ryan Kidd, LauraVaughan, McKennaFitzgerald, Christian Smith, Juan Gil and Henry Sleight

11 May 2024 0:09 UTC

57 points

8 comments49 min readLW link

[Question] How do I get better at D&D Sci?

FinalFormal211 May 2024 18:48 UTC

6 points

0 comments1 min readLW link

Take the wheel, Shoggoth! (LW frontpage algorithm experiments)

Ruby and RobertM

23 Apr 2024 3:58 UTC

62 points

18 comments4 min readLW link

[Question] Resources for learning about poise / gracefulness?

David Gross11 May 2024 18:30 UTC

6 points

0 comments1 min readLW link

Open Thread Spring 2024

habryka11 Mar 2024 19:17 UTC

22 points

119 comments1 min readLW link

ACX Meetups Everywhere Spring 2024, Montreal, QC

BionicD0LPH1N22 Mar 2024 23:15 UTC

6 points

1 comment1 min readLW link

New intro textbook on AIXI

Alex_Altair11 May 2024 18:18 UTC

13 points

0 comments1 min readLW link

Please stop publishing ideas/insights/research about AI

Tamsin Leake2 May 2024 14:54 UTC

12 points

59 comments4 min readLW link

Dating Roundup #3: Third Time’s the Charm

Zvi8 May 2024 13:30 UTC

35 points

25 comments39 min readLW link

(thezvi.wordpress.com)

Should I Finish My Bachelor’s Degree?

Zack_M_Davis11 May 2024 5:17 UTC

12 points

3 comments6 min readLW link

(zackmdavis.net)

ACX Everywhere Norfolk, Virginia

Willa11 May 2024 16:59 UTC

4 points

0 comments1 min readLW link

[Question] Ethics and prospects of AI related jobs?

dr_s11 May 2024 9:31 UTC

10 points

8 comments1 min readLW link

We might be missing some key feature of AI takeoff; it’ll probably seem like “we could’ve seen this coming”

Lukas_Gloor9 May 2024 15:43 UTC

70 points

23 comments5 min readLW link

Chapter 2 The Big Bad Love Machine

David Chapel11 May 2024 9:33 UTC

−4 points

1 comment27 min readLW link

Pascal’s Mugging and the Order of Quantification

Mascal's Pugging10 May 2024 18:34 UTC

10 points

3 comments2 min readLW link

St. Louis – ACX Meetups Everywhere Spring 2024

JohnBuridan30 Mar 2024 11:28 UTC

9 points

3 comments1 min readLW link

Questions are usually too cheap

Nathan Young11 May 2024 13:00 UTC

18 points

4 comments6 min readLW link

(nathanpmyoung.substack.com)