All 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 202220232024 2025

All Jan Feb Mar Apr May Jun Jul Aug Sep Oct NovDec

All 1 2 3 4 5 6 7 8 91011 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31

Conceptual coherence for concrete categories in humans and LLMs

Bill BenzonDec 9, 2023, 11:49 PM

13 points

4 votes

Overall karma indicates overall quality.

1 comment2 min readLW link

2d ai-partners as a comprehensive motivation tool

AiresJLDec 9, 2023, 9:59 PM

3 points

8 votes

Overall karma indicates overall quality.

0 comments1 min readLW link

Without—MicroFiction 250 words

Carissa CassielDec 9, 2023, 9:49 PM

20 points

9 votes

Overall karma indicates overall quality.

1 comment1 min readLW link

Some negative steganography results

Fabien RogerDec 9, 2023, 8:22 PM

60 points

27 votes

Overall karma indicates overall quality.

5 comments2 min readLW link

Summing up “Scheming AIs” (Section 5)

Joe CarlsmithDec 9, 2023, 3:48 PM

2 points

5 votes

Overall karma indicates overall quality.

1 comment11 min readLW link

The Offense-Defense Balance Rarely Changes

Maxwell TabarrokDec 9, 2023, 3:21 PM

77 points

49 votes

Overall karma indicates overall quality.

23 comments3 min readLW link

(maximumprogress.substack.com)

A Philosophical Tautology

Nox MLDec 9, 2023, 2:06 PM

−2 points

7 votes

Overall karma indicates overall quality.

45 comments2 min readLW link

Unpicking Extinction

ukc10014Dec 9, 2023, 9:15 AM

35 points

14 votes

Overall karma indicates overall quality.

10 comments10 min readLW link

Finding Sparse Linear Connections between Features in LLMs

Logan Riggs, Sam Mitchell and Adam Kaufman

Dec 9, 2023, 2:27 AM

70 points

31 votes

Overall karma indicates overall quality.

5 comments10 min readLW link

[Question] Option Space Nomenclature

SilverFlameDec 8, 2023, 11:14 PM

1 point

4 votes

Overall karma indicates overall quality.

0 comments1 min readLW link

“Model UN Solutions”

Arjun PanicksseryDec 8, 2023, 11:06 PM

36 points

21 votes

Overall karma indicates overall quality.

5 comments1 min readLW link

(open.substack.com)

Speed arguments against scheming (Section 4.4-4.7 of “Scheming AIs”)

Joe CarlsmithDec 8, 2023, 9:09 PM

9 points

3 votes

Overall karma indicates overall quality.

0 comments15 min readLW link

Modeling incentives at scale using LLMs

Bruno Marnette, pzahn and cmck

Dec 8, 2023, 6:46 PM

7 points

5 votes

Overall karma indicates overall quality.

3 comments13 min readLW link

Refusal mechanisms: initial experiments with Llama-2-7b-chat

Andy Arditi and Oscar Obeso

Dec 8, 2023, 5:08 PM

82 points

37 votes

Overall karma indicates overall quality.

7 comments7 min readLW link

Colour versus Shape Goal Misgeneralization in Reinforcement Learning: A Case Study

Karolis JucysDec 8, 2023, 1:18 PM

16 points

7 votes

Overall karma indicates overall quality.

1 comment4 min readLW link

(arxiv.org)

What I Would Do If I Were Working On AI Governance

johnswentworthDec 8, 2023, 6:43 AM

112 points

55 votes

Overall karma indicates overall quality.

32 comments10 min readLW link

Whither Prison Abolition?

MadHatterDec 8, 2023, 5:27 AM

−6 points

8 votes

Overall karma indicates overall quality.

0 comments16 min readLW link

(bittertruths.substack.com)

Class consciousness for those against the class system

TekhneMakreDec 8, 2023, 1:02 AM

11 points

11 votes

Overall karma indicates overall quality.

9 comments1 min readLW link

Building selfless agents to avoid instrumental self-preservation.

blalloDec 7, 2023, 6:59 PM

14 points

6 votes

Overall karma indicates overall quality.

2 comments6 min readLW link

Does Chat-GPT display ‘Scope Insensitivity’?

callumDec 7, 2023, 6:58 PM

12 points

5 votes

Overall karma indicates overall quality.

1 comment3 min readLW link

LLM keys—A Proposal of a Solution to Prompt Injection Attacks

Peter HroššoDec 7, 2023, 5:36 PM

1 point

2 votes

Overall karma indicates overall quality.

2 comments1 min readLW link

Meetup Tip: Heartbeat Messages

ScrewtapeDec 7, 2023, 5:18 PM

69 points

32 votes

Overall karma indicates overall quality.

4 comments3 min readLW link

[Valence series] 2. Valence & Normativity

Steven ByrnesDec 7, 2023, 4:43 PM

88 points

28 votes

Overall karma indicates overall quality.

7 comments28 min readLW link 1 review

AISN #27: Defensive Accelerationism, A Retrospective On The OpenAI Board Saga, And A New AI Bill From Senators Thune And Klobuchar

Dan H, Corin Katzke and allison huang

Dec 7, 2023, 3:59 PM

13 points

5 votes

Overall karma indicates overall quality.

0 comments6 min readLW link

(newsletter.safe.ai)

AI #41: Bring in the Other Gemini

ZviDec 7, 2023, 3:10 PM

46 points

29 votes

Overall karma indicates overall quality.

16 comments52 min readLW link

(thezvi.wordpress.com)

Simplicity arguments for scheming (Section 4.3 of “Scheming AIs”)

Joe CarlsmithDec 7, 2023, 3:05 PM

10 points

3 votes

Overall karma indicates overall quality.

1 comment19 min readLW link

Gemini 1.0

ZviDec 7, 2023, 2:40 PM

50 points

34 votes

Overall karma indicates overall quality.

7 comments9 min readLW link

(thezvi.wordpress.com)

Random Musings on Theory of Impact for Activation Vectors

Chris_LeongDec 7, 2023, 1:07 PM

8 points

3 votes

Overall karma indicates overall quality.

0 comments1 min readLW link

[Question] Is AlphaGo actually a consequentialist utility maximizer?

faul_snameDec 7, 2023, 12:41 PM

36 points

13 votes

Overall karma indicates overall quality.

8 comments3 min readLW link

(Report) Evaluating Taiwan’s Tactics to Safeguard its Semiconductor Assets Against a Chinese Invasion

GauraventhDec 7, 2023, 11:50 AM

14 points

10 votes

Overall karma indicates overall quality.

5 comments22 min readLW link

(bristolaisafety.org)

Would AIs trapped in the Metaverse pine to enter the real world and would the ramifications cause trouble?

ProfessorFalkenDec 7, 2023, 10:17 AM

−2 points

3 votes

Overall karma indicates overall quality.

1 comment1 min readLW link

The GiveWiki’s Top Picks in AI Safety for the Giving Season of 2023

Dawn DrescherDec 7, 2023, 9:23 AM

4 points

3 votes

Overall karma indicates overall quality.

10 comments3 min readLW link

(impactmarkets.substack.com)

Language Model Memorization, Copyright Law, and Conditional Pretraining Alignment

RogerDearnaleyDec 7, 2023, 6:14 AM

9 points

5 votes

Overall karma indicates overall quality.

0 comments11 min readLW link

Reflective consistency, randomized decisions, and the dangers of unrealistic thought experiments

Radford NealDec 7, 2023, 3:33 AM

34 points

14 votes

Overall karma indicates overall quality.

25 comments6 min readLW link

[Question] For fun: How long can you hold your breath?

exanovaDec 6, 2023, 11:36 PM

1 point

3 votes

Overall karma indicates overall quality.

7 comments1 min readLW link

Mathematics As Physics

Nox MLDec 6, 2023, 10:27 PM

−2 points

9 votes

Overall karma indicates overall quality.

10 comments5 min readLW link

The counting argument for scheming (Sections 4.1 and 4.2 of “Scheming AIs”)

Joe CarlsmithDec 6, 2023, 7:28 PM

10 points

3 votes

Overall karma indicates overall quality.

0 comments10 min readLW link

On Trust

johnswentworthDec 6, 2023, 7:19 PM

42 points

35 votes

Overall karma indicates overall quality.

26 comments4 min readLW link

Originality vs. Correctness

alkjash and habryka

Dec 6, 2023, 6:51 PM

60 points

20 votes

Overall karma indicates overall quality.

17 comments25 min readLW link

Proposal for improving the global online discourse through personalised comment ordering on all websites

Roman LeventovDec 6, 2023, 6:51 PM

35 points

14 votes

Overall karma indicates overall quality.

21 comments6 min readLW link

Google Gemini Announced

Jacob G-WDec 6, 2023, 4:14 PM

54 points

26 votes

Overall karma indicates overall quality.

22 comments1 min readLW link

(blog.google)

Based Beff Jezos and the Accelerationists

ZviDec 6, 2023, 4:00 PM

90 points

45 votes

Overall karma indicates overall quality.

29 comments12 min readLW link

(thezvi.wordpress.com)

Bucket Brigade: Likely End-of-Life

jefftkDec 6, 2023, 3:30 PM

16 points

5 votes

Overall karma indicates overall quality.

1 comment1 min readLW link

(www.jefftk.com)

Why Yudkowsky is wrong about “covalently bonded equivalents of biology”

titotalDec 6, 2023, 2:09 PM

45 points

96 votes

Overall karma indicates overall quality.

41 comments16 min readLW link

(open.substack.com)

Metaculus Launches Chinese AI Chips Tournament, Supporting Institute for AI Policy and Strategy Research

ChristianWilliamsDec 6, 2023, 11:26 AM

10 points

2 votes

Overall karma indicates overall quality.

1 comment1 min readLW link

(www.metaculus.com)

Minimal Viable Paradise: How do we get The Good Future(TM)?

Nathan YoungDec 6, 2023, 9:24 AM

9 points

5 votes

Overall karma indicates overall quality.

0 comments7 min readLW link

Anthropical Paradoxes are Paradoxes of Probability Theory

Ape in the coatDec 6, 2023, 8:16 AM

55 points

18 votes

Overall karma indicates overall quality.

18 comments5 min readLW link

Digital humans vs merge with AI? Same or different?

Nathan Helm-Burger and mishka

Dec 6, 2023, 4:56 AM

21 points

6 votes

Overall karma indicates overall quality.

11 comments7 min readLW link

EA Infrastructure Fund’s Plan to Focus on Principles-First EA

LinchDec 6, 2023, 3:24 AM

27 points

8 votes

Overall karma indicates overall quality.

0 comments9 min readLW link

In defence of Helen Toner, Adam D’Angelo, and Tasha McCauley

peterrDec 6, 2023, 2:02 AM

25 points

15 votes

Overall karma indicates overall quality.

3 comments9 min readLW link

(pastebin.com)

Keyboard shortcuts

Keys shown in yellow (e.g., ]) are accesskeys, and require a browser-specific modifier key (or keys).

Keys shown in grey (e.g., ?) do not require any modifier keys.

General
? Show keyboard shortcuts
Esc Hide keyboard shortcuts

Site navigation
h Go to Home (a.k.a. “Frontpage”) view
f Go to Featured (a.k.a. “Curated”) view
a Go to All (a.k.a. “Community”) view
m Go to Meta view
v Go to Tags view
c Go to Recent Comments view
r Go to Archive view
q Go to Sequences view
t Go to About page
u Go to User or Login page
o Go to Inbox page

Page navigation
, Jump up to top of page
. Jump down to bottom of page
/ Jump to top of comments section
s Search

Page actions
n New post or comment
e Edit current post

Post/comment list views
. Focus next entry in list
, Focus previous entry in list
; Cycle between links in focused entry
Enter Go to currently focused entry
Esc Unfocus currently focused entry
] Go to next page
[ Go to previous page
\ Go to first page
e Edit currently focused post

Editor
k Bold text
i Italic text
l Insert hyperlink
q Blockquote text

Appearance
= Increase text size
- Decrease text size
0 Reset to default text size
′ Cycle through content width settings
1 Switch to default theme [A]
2 Switch to dark theme [B]
3 Switch to grey theme [C]
4 Switch to ultramodern theme [D]
5 Switch to simple theme [E]
6 Switch to brutalist theme [F]
7 Switch to ReadTheSequences theme [G]
8 Switch to classic Less Wrong theme [H]
9 Switch to modern Less Wrong theme [I]
; Open theme tweaker
Enter Save changes and close theme tweaker
Esc Close theme tweaker (without saving)

Slide shows
l Start/resume slideshow
Esc Exit slideshow
→↓ Next slide
←↑ Previous slide
Space Reset slide zoom

Miscellaneous
x Switch to next view on user page
z Switch to previous view on user page
` Toggle compact comment list view
g Toggle anti-kibitzer