RSS

Has Diagram

TagLast edit: Apr 29, 2023, 10:52 PM by Gunnar_Zarncke

This tag is used to indicate that the post contains diagrams. This may be useful to quickly find such posts, or to exclude them in case you are visually impaired.

What are the re­sults of more parental su­per­vi­sion and less out­door play?

juliawiseNov 25, 2023, 12:52 PM
228 points
31 comments5 min readLW link

Us­ing axis lines for good or evil

dynomightMar 6, 2024, 2:47 PM
150 points
39 comments4 min readLW link
(dynomight.net)

Neu­ral Categories

Eliezer YudkowskyFeb 10, 2008, 12:33 AM
63 points
17 comments4 min readLW link

The lat­tice of par­tial updatelessness

Martín SotoFeb 10, 2024, 5:34 PM
21 points
5 comments5 min readLW link

De­mys­tify­ing “Align­ment” through a Comic

milanroskoJun 9, 2024, 8:24 AM
106 points
19 comments1 min readLW link

Shard The­ory—is it true for hu­mans?

RishikaJun 14, 2024, 7:21 PM
71 points
7 comments15 min readLW link

Towards a Less Bul­lshit Model of Semantics

Jun 17, 2024, 3:51 PM
94 points
44 comments21 min readLW link

How good are LLMs at do­ing ML on an un­known dataset?

Håvard Tveit IhleJul 1, 2024, 9:04 AM
33 points
4 comments13 min readLW link

[In­tro to brain-like-AGI safety] 4. The “short-term pre­dic­tor”

Steven ByrnesFeb 16, 2022, 1:12 PM
64 points
11 comments13 min readLW link

An In­tro­duc­tion To The Man­delbrot Set That Doesn’t Men­tion Com­plex Numbers

YitzJan 17, 2024, 9:48 AM
82 points
11 comments9 min readLW link

An Illus­trated Proof of the No Free Lunch Theorem

lifelonglearnerJun 8, 2020, 1:54 AM
19 points
0 comments1 min readLW link
(mlu.red)

Cor­rigi­bil­ity, Much more de­tail than any­one wants to Read

Logan ZoellnerMay 7, 2023, 1:02 AM
26 points
2 comments7 min readLW link

How much do you be­lieve your re­sults?

Eric NeymanMay 6, 2023, 8:31 PM
496 points
18 comments15 min readLW link4 reviews
(ericneyman.wordpress.com)

Resi­d­ual stream norms grow ex­po­nen­tially over the for­ward pass

May 7, 2023, 12:46 AM
77 points
24 comments11 min readLW link

Be­ing the (Pareto) Best in the World

johnswentworthJun 24, 2019, 6:36 PM
466 points
60 comments3 min readLW link3 reviews

Hyperpolation

Gunnar_ZarnckeSep 15, 2024, 9:37 PM
22 points
6 comments1 min readLW link
(arxiv.org)

The case for a nega­tive al­ign­ment tax

Sep 18, 2024, 6:33 PM
75 points
20 comments7 min readLW link

Ma­chine Learn­ing Anal­ogy for Med­i­ta­tion (illus­trated)

abramdemskiJun 28, 2018, 10:51 PM
100 points
48 comments1 min readLW link

How might we solve the al­ign­ment prob­lem? (Part 1: In­tro, sum­mary, on­tol­ogy)

Joe CarlsmithOct 28, 2024, 9:57 PM
54 points
5 comments32 min readLW link

I turned de­ci­sion the­ory prob­lems into memes about trolleys

TapataktOct 30, 2024, 8:13 PM
104 points
23 comments1 min readLW link

The Car­toon Guide to Löb’s Theorem

Eliezer YudkowskyAug 17, 2008, 8:35 PM
44 points
104 comments1 min readLW link

[In­tro to brain-like-AGI safety] 10. The al­ign­ment problem

Steven ByrnesMar 30, 2022, 1:24 PM
48 points
7 comments19 min readLW link

[In­tro to brain-like-AGI safety] 12. Two paths for­ward: “Con­trol­led AGI” and “So­cial-in­stinct AGI”

Steven ByrnesApr 20, 2022, 12:58 PM
44 points
10 comments15 min readLW link

Draw­ing Less Wrong: Tech­ni­cal Skill

RaemonDec 5, 2011, 5:12 AM
37 points
36 comments9 min readLW link

All images from the WaitButWhy se­quence on AI

trevorApr 8, 2023, 7:36 AM
73 points
5 comments2 min readLW link

The Nat­u­ral Ab­strac­tion Hy­poth­e­sis: Im­pli­ca­tions and Evidence

CallumMcDougallDec 14, 2021, 11:14 PM
39 points
9 comments19 min readLW link

Test­ing The Nat­u­ral Ab­strac­tion Hy­poth­e­sis: Pro­ject Update

johnswentworthSep 20, 2021, 3:44 AM
88 points
17 comments8 min readLW link1 review

Open tech­ni­cal prob­lem: A Quinean proof of Löb’s the­o­rem, for an eas­ier car­toon guide

Andrew_CritchNov 24, 2022, 9:16 PM
58 points
35 comments3 min readLW link1 review

[In­tro to brain-like-AGI safety] 5. The “long-term pre­dic­tor”, and TD learning

Steven ByrnesFeb 23, 2022, 2:44 PM
54 points
27 comments20 min readLW link

[In­tro to brain-like-AGI safety] 6. Big pic­ture of mo­ti­va­tion, de­ci­sion-mak­ing, and RL

Steven ByrnesMar 2, 2022, 3:26 PM
69 points
17 comments16 min readLW link

[In­tro to brain-like-AGI safety] 7. From hard­coded drives to fore­sighted plans: A worked example

Steven ByrnesMar 9, 2022, 2:28 PM
78 points
0 comments10 min readLW link

[In­tro to brain-like-AGI safety] 8. Take­aways from neuro 1/​2: On AGI development

Steven ByrnesMar 16, 2022, 1:59 PM
57 points
2 comments14 min readLW link

[In­tro to brain-like-AGI safety] 9. Take­aways from neuro 2/​2: On AGI motivation

Steven ByrnesMar 23, 2022, 12:48 PM
46 points
11 comments22 min readLW link

[In­tro to brain-like-AGI safety] 13. Sym­bol ground­ing & hu­man so­cial instincts

Steven ByrnesApr 27, 2022, 1:30 PM
73 points
15 comments15 min readLW link

[In­tro to brain-like-AGI safety] 14. Con­trol­led AGI

Steven ByrnesMay 11, 2022, 1:17 PM
45 points
25 comments20 min readLW link

[In­tro to brain-like-AGI safety] 1. What’s the prob­lem & Why work on it now?

Steven ByrnesJan 26, 2022, 3:23 PM
158 points
19 comments26 min readLW link

[In­tro to brain-like-AGI safety] 2. “Learn­ing from scratch” in the brain

Steven ByrnesFeb 2, 2022, 1:22 PM
59 points
12 comments25 min readLW link

[In­tro to brain-like-AGI safety] 3. Two sub­sys­tems: Learn­ing & Steering

Steven ByrnesFeb 9, 2022, 1:09 PM
95 points
3 comments25 min readLW link

[Valence se­ries] 4. Valence & So­cial Sta­tus (de­p­re­cated)

Steven ByrnesDec 15, 2023, 2:24 PM
35 points
19 comments11 min readLW link

Bayes’ The­o­rem Illus­trated (My Way)

komponistoJun 3, 2010, 4:40 AM
171 points
195 comments9 min readLW link

In­duc­tion heads—illustrated

CallumMcDougallJan 2, 2023, 3:35 PM
125 points
10 comments3 min readLW link

Vi­su­al­iz­ing small At­ten­tion-only Transformers

WCargoNov 19, 2024, 9:37 AM
4 points
0 comments8 min readLW link

Levels of goals and alignment

zeshenSep 16, 2022, 4:44 PM
27 points
4 comments6 min readLW link

A new­comer’s guide to the tech­ni­cal AI safety field

zeshenNov 4, 2022, 2:29 PM
42 points
3 comments10 min readLW link

Embed­ding safety in ML development

zeshenOct 31, 2022, 12:27 PM
24 points
1 comment18 min readLW link