This website requires javascript to properly function. Consider activating javascript to get access to all site functionality.
AI ALIGNMENT FORUM
AF
Login
Connor Kissane
Posts
Sorted by New
35
Base LLMs refuse too
8d
10
26
SAEs (usually) Transfer Between Base and Chat Models
3mo
0
18
Attention Output SAEs Improve Circuit Analysis
4mo
0
33
We Inspected Every Head In GPT-2 Small using SAEs So You Don’t Have To
7mo
0
28
Attention SAEs Scale to GPT-2 Small
8mo
0
35
Sparse Autoencoders Work on Attention Layer Outputs
9mo
3
Wiki Contributions
Comments
Sorted by
Newest