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
34
SAEs are highly dataset dependent: a case study on the refusal direction
7mo
0
23
Open Source Replication of Anthropic’s Crosscoder paper for model-diffing
7mo
3
35
Base LLMs refuse too
8mo
10
28
SAEs (usually) Transfer Between Base and Chat Models
11mo
0
18
Attention Output SAEs Improve Circuit Analysis
1y
0
34
We Inspected Every Head In GPT-2 Small using SAEs So You Don’t Have To
1y
0
28
Attention SAEs Scale to GPT-2 Small
1y
0
35
Sparse Autoencoders Work on Attention Layer Outputs
1y
3
Wikitag Contributions
Comments
Sorted by
Newest