AI ALIGNMENT FORUMTags
AF

Deconfusion

EditHistory
Discussion (0)
Help improve this page (3 flags)
EditHistory
Discussion (0)
Help improve this page (3 flags)
Deconfusion
Random Tag
Contributors
2Abram Demski

Narrowly, deconfusion is a specific branch of AI alignment research, discussed in MIRI's 2018 research update. More broadly, the term applies to any domain. Quoting from the research update:

By deconfusion, I mean something like “making it so that you can think about a given topic without continuously accidentally spouting nonsense.”

Posts tagged Deconfusion
5
26Looking Deeper at Deconfusion
Adam Shimi
2y
2
1
30Deconfusing Direct vs Amortised Optimization
Beren Millidge
6mo
3
1
35Modelling Transformative AI Risks (MTAIR) Project: Introduction
David Manheim, Aryeh Englander
2y
0
1
43Builder/Breaker for Deconfusion
Abram Demski
9mo
8
1
24Applications for Deconfusing Goal-Directedness
Adam Shimi
2y
3
1
21Musings on general systems alignment
Alex Flint
2y
1
1
15Traps of Formalization in Deconfusion
Adam Shimi
2y
2
1
15Open problem: how can we quantify player alignment in 2x2 normal-form games?Q
Alex Turner, Vanessa Kosoy
2y
Q
21
1
10Approaches to gradient hacking
Adam Shimi
2y
7
1
12Alex Turner's Research, Comprehensive Information Gathering
Adam Shimi
2y
3
1
9Goal-Directedness and Behavior, Redux
Adam Shimi
2y
2
1
8A review of "Agents and Devices"
Adam Shimi
2y
0
1
10Power-seeking for successive choices
Adam Shimi
2y
9
1
139Simulators
janus
9mo
62
0
85Reward is not the optimization target
Alex Turner
1y
80
Load More (15/19)
Add Posts