WISDOMISM A Moral Theory for the Age of Information

Peter lawless 19 Apr 2024 23:06 UTC

4 points

0 comments9 min readLW link

Inducing Unprompted Misalignment in LLMs

Sam Svenningsen, evhub and Henry Sleight

19 Apr 2024 20:00 UTC

22 points

5 comments16 min readLW link

Progress Update #1 from the GDM Mech Interp Team: Full Update

Neel Nanda, Arthur Conmy, lsgos, Senthooran Rajamanoharan, Tom Lieberum, János Kramár and Vikrant Varma

19 Apr 2024 19:06 UTC

36 points

4 comments8 min readLW link

Progress Update #1 from the GDM Mech Interp Team: Summary

Neel Nanda, Arthur Conmy, lsgos, Senthooran Rajamanoharan, Tom Lieberum, János Kramár and Vikrant Varma

19 Apr 2024 19:06 UTC

32 points

0 comments3 min readLW link

[Question] What is the best way to talk about probabilities you expect to change with evidence/experiments?

Will_Pearson19 Apr 2024 15:35 UTC

12 points

9 comments1 min readLW link

CTMU insight: maybe consciousness can affect quantum outcomes?

zhukeepa19 Apr 2024 15:23 UTC

12 points

3 comments4 min readLW link

[Question] How to Model the Future of Open-Source LLMs?

Joel Burget19 Apr 2024 14:28 UTC

10 points

1 comment1 min readLW link

What’s up with all the non-Mormons? Weirdly specific universalities across LLMs

mwatkins19 Apr 2024 13:43 UTC

24 points

4 comments27 min readLW link

[Question] If digital goods in virtual worlds increase GDP, do we actually become richer?

No77e19 Apr 2024 10:06 UTC

4 points

6 comments1 min readLW link

Experiment on repeating choices

KatjaGrace19 Apr 2024 4:20 UTC

50 points

1 comment3 min readLW link

(worldspiritsockpuppet.com)

Cohesion and business problems

Adam Zerner19 Apr 2024 0:45 UTC

10 points

2 comments4 min readLW link

The Thermodynamics of Death

Peter lawless 19 Apr 2024 0:36 UTC

4 points

0 comments10 min readLW link

hydrogen tube transport

bhauth18 Apr 2024 22:47 UTC

26 points

7 comments5 min readLW link

(www.bhauth.com)

A Review of In-Context Learning Hypotheses for Automated AI Alignment Research

alamerton18 Apr 2024 18:29 UTC

20 points

4 comments15 min readLW link

Blessed information, garbage information, cursed information

tailcalled18 Apr 2024 16:56 UTC

20 points

5 comments3 min readLW link

[Fiction] A Confession

Arjun Panickssery18 Apr 2024 16:28 UTC

28 points

2 comments5 min readLW link

(arjunpanickssery.substack.com)

Discriminating Behaviorally Identical Classifiers: a model problem for applying interpretability to scalable oversight

Sam Marks18 Apr 2024 16:17 UTC

75 points

1 comment12 min readLW link

Cooperation is optimal, with weaker agents too - tldr

Ryo 18 Apr 2024 15:03 UTC

12 points

16 comments4 min readLW link

(medium.com)

How to coordinate despite our biases? - tldr

Ryo 18 Apr 2024 15:03 UTC

3 points

2 comments3 min readLW link

(medium.com)

UDT1.01: Logical Inductors and Implicit Beliefs (5/10)

Diffractor18 Apr 2024 8:39 UTC

28 points

1 comment19 min readLW link