22 Jul 2023 20:21 UTC

21 points

5 comments6 min readLW link

Think like a consultant not a salesperson

Adam Zerner22 Jul 2023 19:31 UTC

16 points

5 comments2 min readLW link

Optimization, loss set at variance in RL

Clairstan22 Jul 2023 18:25 UTC

1 point

1 comment3 min readLW link

Compute Thresholds: proposed rules to mitigate risk of a “lab leak” accident during AI training runs

davidad22 Jul 2023 18:09 UTC

80 points

2 comments2 min readLW link

Apollo Neuro Follow Up

Elizabeth22 Jul 2023 17:20 UTC

28 points

0 comments1 min readLW link

(acesounderglass.com)

Expert trap – Ways out (Part 3 of 3)

Paweł Sysiak22 Jul 2023 13:06 UTC

4 points

0 comments9 min readLW link

GPTs’ ability to keep a secret is weirdly prompt-dependent

Mateusz Bagiński, Filip Sondej and Marcel Windys

22 Jul 2023 12:21 UTC

31 points

0 comments9 min readLW link

Replacing the Big Air Purifier

jefftk22 Jul 2023 12:10 UTC

10 points

0 comments1 min readLW link

(www.jefftk.com)

[Question] I’m consistently overwhelmed by basic obligations. Are there any paradigm shifts or other rationality-based tips that would be helpful?

Benjamin Hendricks21 Jul 2023 21:10 UTC

60 points

37 comments2 min readLW link

Fundamentally Fuzzy Concepts Can’t Have Crisp Definitions: Cooperation and Alignment vs Math and Physics

VojtaKovarik21 Jul 2023 21:03 UTC

12 points

18 comments3 min readLW link

Cooking Air Quality

jefftk21 Jul 2023 19:30 UTC

16 points

1 comment2 min readLW link

(www.jefftk.com)

Reward Hacking from a Causal Perspective

tom4everitt, Francis Rhys Ward, sbenthall, James Fox, mattmacdermott and RyanCarey

21 Jul 2023 18:27 UTC

29 points

5 comments7 min readLW link

News : Biden-⁠Harris Administration Secures Voluntary Commitments from Leading Artificial Intelligence Companies to Manage the Risks Posed by AI

Jonathan Claybrough21 Jul 2023 18:00 UTC

65 points

9 comments2 min readLW link

(www.whitehouse.gov)

The UAP Disclosure Act of 2023 and its implications

andeslodes21 Jul 2023 17:21 UTC

36 points

47 comments20 min readLW link

(www.congress.gov)

To use computers well, learn their rules

dkl921 Jul 2023 17:00 UTC

4 points

6 comments4 min readLW link

(dkl9.net)

BCIs and the ecosystem of modular minds

beren21 Jul 2023 15:58 UTC

84 points

14 comments11 min readLW link

Priorities for the UK Foundation Models Taskforce

Andrea_Miotti21 Jul 2023 15:23 UTC

105 points

4 comments5 min readLW link

(www.conjecture.dev)

Training Process Transparency through Gradient Interpretability: Early experiments on toy language models

robertzk and evhub

21 Jul 2023 14:52 UTC

56 points

1 comment1 min readLW link

[Question] Can AI Alignment please create a Reddit-like platform that would make it much easier for alignment researchers to find and help each other?

Georgeo5721 Jul 2023 14:03 UTC

−5 points

2 comments1 min readLW link

Case for Foundation Models beyond English

Varshul Gupta21 Jul 2023 13:59 UTC

1 point

0 comments3 min readLW link

(dubverseblack.substack.com)

Meta is hiring for LLM red teaming position

Michael Tontchev21 Jul 2023 13:57 UTC

7 points

0 comments1 min readLW link

(us.meta.talentnet.community)

[Linkpost] Interpreting Multimodal Video Transformers Using Brain Recordings

Bogdan Ionut Cirstea21 Jul 2023 11:26 UTC

5 points

0 comments1 min readLW link

Berlin AI Alignment Open Meetup August 2023

GuyP21 Jul 2023 10:58 UTC

1 point

0 comments1 min readLW link

Decoding intermediate activations in llama-2-7b

Nina Rimsky21 Jul 2023 5:35 UTC

36 points

3 comments4 min readLW link

GPT-2′s positional embedding matrix is a helix

AdamYedidia21 Jul 2023 4:16 UTC

42 points

18 comments4 min readLW link

Problems with predictive history classes

dkl920 Jul 2023 23:28 UTC

15 points

5 comments1 min readLW link

Announcement: AI Narrations Available for All New LessWrong Posts

Solenoid_Entity, Ruby, Raemon, PeterH and TYPE III AUDIO

20 Jul 2023 22:17 UTC

67 points

28 comments1 min readLW link

AI #21: The Cup Overfloweth

Zvi20 Jul 2023 21:30 UTC

47 points

4 comments64 min readLW link

(thezvi.wordpress.com)

All AGI Safety questions welcome (especially basic ones) [July 2023]

smallsilo20 Jul 2023 20:20 UTC

38 points

40 comments2 min readLW link

(forum.effectivealtruism.org)

Growth of Publicly Available Genetic Sequencing Data

jefftk20 Jul 2023 19:50 UTC

11 points

2 comments1 min readLW link

(www.jefftk.com)

Progress links and tweets, 2023-07-20: “A goddess enthroned on a car”

jasoncrawford20 Jul 2023 18:28 UTC

12 points

4 comments2 min readLW link

(rootsofprogress.org)

Boundary Placement Rebellion

tailcalled20 Jul 2023 17:40 UTC

53 points

21 comments12 min readLW link

Going Beyond Linear Mode Connectivity: The Layerwise Linear Feature Connectivity

zhanpeng_zhou20 Jul 2023 17:38 UTC

22 points

13 comments3 min readLW link

(openreview.net)

Even Superhuman Go AIs Have Surprising Failure Modes

AdamGleave, EuanMcLean, Tony Wang, Kellin Pelrine, Tom Tseng, Yawen Duan, Joseph Miller and MichaelDennis

20 Jul 2023 17:31 UTC

126 points

21 comments10 min readLW link

(far.ai)

Paper digestion: “May We Have Your Attention Please? Human-Rights NGOs and the Problem of Global Communication”

Klara Helene Nielsen20 Jul 2023 17:08 UTC

4 points

1 comment2 min readLW link

(journals.sagepub.com)

The (short) case for predicting what Aliens value

Jim Buhler20 Jul 2023 15:25 UTC

11 points

5 comments3 min readLW link

Does Circuit Analysis Interpretability Scale? Evidence from Multiple Choice Capabilities in Chinchilla

Neel Nanda, Tom Lieberum, Matthew Rahtz, János Kramár, Geoffrey Irving, Rohin Shah and Vlad Mikulik

20 Jul 2023 10:50 UTC

44 points

3 comments2 min readLW link

(arxiv.org)

Speculative inferences about path dependence in LLM supervised fine-tuning from results on linear mode connectivity and model souping

RobertKirk20 Jul 2023 9:56 UTC

38 points

2 comments5 min readLW link

A case for gamete personhood (reductio ad absurdum)

Ansyn131220 Jul 2023 8:25 UTC

−1 points

4 comments1 min readLW link

Contra Contra the Social Model of Disability

DirectedEvolution20 Jul 2023 6:59 UTC

20 points

22 comments16 min readLW link

[Question] Do you speed up capabilities when you do AI integrations and consume overhangs?

Michael Tontchev20 Jul 2023 6:40 UTC

6 points

1 comment1 min readLW link

[Question] How necessary is intuition, for advanced math?

NicholasKross20 Jul 2023 0:18 UTC

11 points

8 comments1 min readLW link

Project Lawful Audiobook: An Unofficial Fan Production with ElevenLabs AI

Askwho19 Jul 2023 23:34 UTC

17 points

1 comment1 min readLW link

(askwhocastsai.substack.com)

Using predictors in corrigible systems

porby19 Jul 2023 22:29 UTC

19 points

6 comments27 min readLW link

mental number lines

bhauth19 Jul 2023 21:01 UTC

10 points

5 comments1 min readLW link

[Question] Any suggestions for an impactful master’s thesis in Political Science?

Klara Helene Nielsen19 Jul 2023 17:44 UTC

1 point

0 comments1 min readLW link

Incident reporting for AI safety

Zach Stein-Perlman19 Jul 2023 17:00 UTC

22 points

0 comments1 min readLW link

Alignment Grantmaking is Funding-Limited Right Now

johnswentworth19 Jul 2023 16:49 UTC

307 points

67 comments1 min readLW link

Zener Science

Screwtape19 Jul 2023 16:40 UTC

16 points

11 comments6 min readLW link

Tallinn, Estonia ACX Summer Meetup

Andrew19 Jul 2023 16:22 UTC

1 point

1 comment1 min readLW link