Go Back
Choose this branch
Choose this branch
meritocratic
regular
democratic
hot
top
alive
49 posts
Conjecture (org)
Refine
Project Announcement
Encultured AI (org)
63 posts
Language Models
Anthropic
Exploratory Engineering
Transformer Circuits
Transformers
101
[Interim research report] Taking features out of superposition with sparse autoencoders
Lee Sharkey
7d
10
234
Conjecture: a retrospective after 8 months of work
Connor Leahy
27d
9
262
Mysteries of mode collapse
janus
1mo
35
103
Searching for Search
NicholasKees
22d
6
125
Conjecture Second Hiring Round
Connor Leahy
27d
0
32
Tradeoffs in complexity, abstraction, and generality
remember
8d
0
103
What I Learned Running Refine
adamShimi
26d
5
130
Current themes in mechanistic interpretability research
Lee Sharkey
1mo
3
68
The First Filter
adamShimi
24d
5
44
Good Futures Initiative: Winter Project Internship
Aris
23d
4
234
Connor Leahy on Dying with Dignity, EleutherAI and Conjecture
Michaƫl Trazzi
5mo
29
105
Understanding Conjecture: Notes from Connor Leahy interview
Akash
3mo
24
267
We Are Conjecture, A New Alignment Research Startup
Connor Leahy
8mo
24
5
Creating a database for base rates
nikos
8d
1
27
Discovering Language Model Behaviors with Model-Written Evaluations
evhub
4h
3
32
Take 11: "Aligning language models" should be weirder.
Charlie Steiner
2d
0
7
Will research in AI risk jinx it? Consequences of training AI on AI risk arguments
Yann Dubois
1d
6
52
Discovering Latent Knowledge in Language Models Without Supervision
Xodarap
6d
1
148
Did ChatGPT just gaslight me?
ThomasW
19d
45
808
Simulators
janus
3mo
103
36
An exploration of GPT-2's embedding weights
Adam Scherlis
7d
2
37
A brainteaser for language models
Adam Scherlis
8d
3
21
Shh, don't tell the AI it's likely to be evil
naterush
14d
9
37
Gliders in Language Models
Alexandre Variengien
25d
11
173
Language models seem to be much better than humans at next-token prediction
Buck
4mo
56
105
Inverse Scaling Prize: Round 1 Winners
Ethan Perez
2mo
16
9
Does a LLM have a utility function?
Dagon
11d
6
61
They gave LLMs access to physics simulators
ryan_b
2mo
18