Go Back
You can't go any further
You can't go any further
meritocratic
regular
democratic
hot
top
alive
2 posts
Distributional Shifts
26 posts
SERI MATS
7
Distribution Shifts and The Importance of AI Safety
Leon Lang
2mo
2
-1
Mesa-optimization for goals defined only within a training environment is dangerous
Rubi J. Hudson
4mo
2
96
Taking the parameters which seem to matter and rotating them until they don't
Garrett Baker
3mo
48
72
Externalized reasoning oversight: a research direction for language model alignment
tamera
4mo
22
61
SERI MATS Program - Winter 2022 Cohort
Ryan Kidd
2mo
12
54
Neural Tangent Kernel Distillation
Thomas Larsen
2mo
20
39
Proper scoring rules don’t guarantee predicting fixed points
Johannes_Treutlein
4d
2
36
Race Along Rashomon Ridge
Stephen Fowler
5mo
15
30
Framing AI Childhoods
David Udell
3mo
8
28
The Ground Truth Problem (Or, Why Evaluating Interpretability Methods Is Hard)
Jessica Mary
1mo
2
28
Behaviour Manifolds and the Hessian of the Total Loss - Notes and Criticism
Spencer Becker-Kahn
3mo
4
27
Finding Skeletons on Rashomon Ridge
David Udell
4mo
2
21
Broad Basins and Data Compression
Jeremy Gillen
4mo
6
20
Why I'm Working On Model Agnostic Interpretability
Jessica Mary
1mo
9
18
Modelling Deception
Garrett Baker
5mo
0
17
Project proposal: Testing the IBP definition of agent
Jeremy Gillen
4mo
4