Go Back
Choose this branch
Choose this branch
meritocratic
regular
democratic
hot
top
alive
55 posts
Language Models
Agency
Deconfusion
Scaling Laws
Tool AI
Definitions
Simulation Hypothesis
PaLM
Prompt Engineering
Philosophy of Language
Carving / Clustering Reality
Astronomical Waste
33 posts
Conjecture (org)
Refine
Project Announcement
Encultured AI (org)
Analogy
234
chinchilla's wild implications
nostalgebraist
4mo
114
185
Simulators
janus
3mo
103
163
Language models seem to be much better than humans at next-token prediction
Buck
4mo
56
157
Transformer Circuits
evhub
12mo
4
141
Announcing the Inverse Scaling Prize ($250k Prize Pool)
Ethan Perez
5mo
14
113
The case for becoming a black-box investigator of language models
Buck
7mo
19
108
Beyond Astronomical Waste
Wei_Dai
4y
41
106
Testing PaLM prompts on GPT3
Yitz
8mo
15
99
Help ARC evaluate capabilities of current language models (still need people)
Beth Barnes
5mo
6
84
Who models the models that model models? An exploration of GPT-3's in-context model fitting ability
Lovre
6mo
14
78
Thoughts on the Alignment Implications of Scaling Language Models
leogao
1y
11
75
Inverse Scaling Prize: Round 1 Winners
Ethan Perez
2mo
16
65
Causal confusion as an argument against the scaling hypothesis
RobertKirk
6mo
30
64
RL with KL penalties is better seen as Bayesian inference
Tomek Korbak
6mo
15
178
Mysteries of mode collapse
janus
1mo
35
143
Conjecture: a retrospective after 8 months of work
Connor Leahy
27d
9
118
We Are Conjecture, A New Alignment Research Startup
Connor Leahy
8mo
24
108
What I Learned Running Refine
adamShimi
26d
5
105
Announcing Encultured AI: Building a Video Game
Andrew_Critch
4mo
26
96
The Singular Value Decompositions of Transformer Weight Matrices are Highly Interpretable
beren
22d
27
76
Refine: An Incubator for Conceptual Alignment Research Bets
adamShimi
8mo
13
70
Announcing the Vitalik Buterin Fellowships in AI Existential Safety!
DanielFilan
1y
2
68
How to Diversify Conceptual Alignment: the Model Behind Refine
adamShimi
5mo
11
64
[Interim research report] Taking features out of superposition with sparse autoencoders
Lee Sharkey
7d
10
61
Circumventing interpretability: How to defeat mind-readers
Lee Sharkey
5mo
8
56
Interpreting Neural Networks through the Polytope Lens
Sid Black
2mo
26
52
Conjecture Second Hiring Round
Connor Leahy
27d
0
48
I missed the crux of the alignment problem the whole time
zeshen
4mo
7