Go Back
Choose this branch
Choose this branch
meritocratic
regular
democratic
hot
top
alive
41 posts
Reinforcement Learning
AI Capabilities
Wireheading
Reward Functions
EfficientZero
Tradeoffs
33 posts
Embedded Agency
Subagents
Robust Agents
Category Theory
Spurious Counterfactuals
Memetics
Autonomous Vehicles
273
EfficientZero: How It Works
1a3orn
1y
42
252
Reward is not the optimization target
TurnTrout
4mo
97
134
EfficientZero: human ALE sample-efficiency w/MuZero+self-supervised
gwern
1y
52
114
We have achieved Noob Gains in AI
phdead
7mo
21
94
Evaluations project @ ARC is hiring a researcher and a webdev/engineer
Beth Barnes
3mo
7
87
The alignment problem in different capability regimes
Buck
1y
12
86
Scaling Laws for Reward Model Overoptimization
leogao
2mo
11
82
Jitters No Evidence of Stupidity in RL
1a3orn
1y
18
77
OpenAI Solves (Some) Formal Math Olympiad Problems
Michaël Trazzi
10mo
26
76
Seriously, what goes wrong with "reward the agent when it makes you smile"?
TurnTrout
4mo
41
74
Will we run out of ML data? Evidence from projecting dataset size trends
Pablo Villalobos
1mo
12
69
Towards deconfusing wireheading and reward maximization
leogao
3mo
7
59
Big picture of phasic dopamine
Steven Byrnes
1y
18
59
My take on Michael Littman on "The HCI of HAI"
Alex Flint
1y
4
248
Humans are very reliable agents
alyssavance
6mo
35
161
Why Subagents?
johnswentworth
3y
42
145
Introduction to Cartesian Frames
Scott Garrabrant
2y
29
143
Embedded Agency (full-text version)
Scott Garrabrant
4y
15
110
Robust Delegation
abramdemski
4y
10
105
Reward Is Not Enough
Steven Byrnes
1y
18
100
Subsystem Alignment
abramdemski
4y
12
88
Embedded Curiosities
Scott Garrabrant
4y
1
87
Embedded World-Models
abramdemski
4y
16
79
Humans Are Embedded Agents Too
johnswentworth
2y
19
73
Updates and additions to "Embedded Agency"
Rob Bensinger
2y
1
62
(A -> B) -> A
Scott Garrabrant
4y
11
61
Additive Operations on Cartesian Frames
Scott Garrabrant
2y
6
50
Functors and Coarse Worlds
Scott Garrabrant
2y
4