Go Back
You can't go any further
You can't go any further
meritocratic
regular
democratic
hot
top
alive
22 posts
Conjecture (org)
12 posts
Refine
59
[Interim research report] Taking features out of superposition with sparse autoencoders
Lee Sharkey
7d
10
132
Conjecture: a retrospective after 8 months of work
Connor Leahy
27d
9
103
What I Learned Running Refine
adamShimi
26d
5
164
Mysteries of mode collapse
janus
1mo
35
22
Tradeoffs in complexity, abstraction, and generality
remember
8d
0
42
The First Filter
adamShimi
24d
5
45
Conjecture Second Hiring Round
Connor Leahy
27d
0
25
Searching for Search
NicholasKees
22d
6
34
Current themes in mechanistic interpretability research
Lee Sharkey
1mo
3
101
Understanding Conjecture: Notes from Connor Leahy interview
Akash
3mo
24
118
Connor Leahy on Dying with Dignity, EleutherAI and Conjecture
Michaël Trazzi
5mo
29
63
How to Diversify Conceptual Alignment: the Model Behind Refine
adamShimi
5mo
11
105
We Are Conjecture, A New Alignment Research Startup
Connor Leahy
8mo
24
43
Refine's First Blog Post Day
adamShimi
4mo
3
22
Embedding safety in ML development
zeshen
1mo
1
45
I missed the crux of the alignment problem the whole time
zeshen
4mo
7
22
confusion about alignment requirements
carado
2mo
10
24
(Structural) Stability of Coupled Optimizers
Paul Bricman
2mo
0
39
the Insulated Goal-Program idea
carado
4mo
3
23
Representational Tethers: Tying AI Latents To Human Ones
Paul Bricman
3mo
0
32
All the posts I will never write
Alexander Gietelink Oldenziel
4mo
8
23
goal-program bricks
carado
4mo
2
12
Refine's Third Blog Post Day/Week
adamShimi
3mo
0
9
Boolean Primitives for Coupled Optimizers
Paul Bricman
2mo
0
11
Refine Blogpost Day #3: The shortforms I did write
Alexander Gietelink Oldenziel
3mo
0
12
Benchmarking Proposals on Risk Scenarios
Paul Bricman
4mo
2