Go Back
Choose this branch
Choose this branch
meritocratic
regular
democratic
hot
top
alive
246 posts
Interpretability (ML & AI)
Instrumental Convergence
Corrigibility
Myopia
Deconfusion
Self Fulfilling/Refuting Prophecies
Orthogonality Thesis
AI Success Models
Treacherous Turn
Gradient Hacking
Mild Optimization
Quantilization
112 posts
Iterated Amplification
Debate (AI safety technique)
Factored Cognition
Humans Consulting HCH
Experiments
Ought
AI-assisted Alignment
Memory and Mnemonics
Air Conditioning
Verification
338
A Mechanistic Interpretability Analysis of Grokking
Neel Nanda
4mo
39
211
The Plan - 2022 Update
johnswentworth
19d
33
205
Debate on Instrumental Convergence between LeCun, Russell, Bengio, Zador, and More
Ben Pace
3y
60
170
Sorting Pebbles Into Correct Heaps
Eliezer Yudkowsky
14y
109
159
The Singular Value Decompositions of Transformer Weight Matrices are Highly Interpretable
beren
22d
27
153
Seeking Power is Often Convergently Instrumental in MDPs
TurnTrout
3y
38
144
Self-fulfilling correlations
PhilGoetz
12y
50
139
Opinions on Interpretable Machine Learning and 70 Summaries of Recent Papers
lifelonglearner
1y
16
136
MIRI comments on Cotra's "Case for Aligning Narrowly Superhuman Models"
Rob Bensinger
1y
13
136
A transparency and interpretability tech tree
evhub
6mo
10
135
Understanding “Deep Double Descent”
evhub
3y
51
123
How "Discovering Latent Knowledge in Language Models Without Supervision" Fits Into a Broader Alignment Scheme
Collin
5d
18
123
Interpreting Neural Networks through the Polytope Lens
Sid Black
2mo
26
118
Deep Learning Systems Are Not Less Interpretable Than Logic/Probability/Etc
johnswentworth
6mo
52
151
Godzilla Strategies
johnswentworth
6mo
65
125
Paul's research agenda FAQ
zhukeepa
4y
73
125
Debate update: Obfuscated arguments problem
Beth Barnes
1y
21
119
My Understanding of Paul Christiano's Iterated Amplification AI Safety Research Agenda
Chi Nguyen
2y
21
118
Supervise Process, not Outcomes
stuhlmueller
8mo
8
109
Preregistration: Air Conditioner Test
johnswentworth
8mo
64
98
Solving Math Problems by Relay
bgold
2y
26
94
Writeup: Progress on AI Safety via Debate
Beth Barnes
2y
18
92
Beliefs and Disagreements about Automating Alignment Research
Ian McKenzie
3mo
4
92
Imitative Generalisation (AKA 'Learning the Prior')
Beth Barnes
1y
14
87
Ought: why it matters and ways to help
paulfchristiano
3y
7
80
Air Conditioner Test Results & Discussion
johnswentworth
6mo
38
74
Model splintering: moving from one imperfect model to another
Stuart_Armstrong
2y
10
73
Why I'm excited about Debate
Richard_Ngo
1y
12