SAELens
ArthurConmyGDM commited on
Commit
fea284f
1 Parent(s): 2a2c10c

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +38 -3
README.md CHANGED
@@ -1,3 +1,38 @@
1
- ---
2
- license: apache-2.0
3
- ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ ---
4
+
5
+ # 1. GemmaScope
6
+
7
+ Gemmascope is TODO
8
+
9
+ # 2. What Is `gemmascope-9b-pt-mlp`?
10
+
11
+ - `gemmascope-`: see 1.
12
+ - `9b-pt-`: These SAEs were trained on the Gemma v2 9B base model (TODO link)
13
+ - `mlp`: These SAEs were trained on the outputs of MLP layers.
14
+
15
+ ## 3. GTM FAQ (TODO(conmy): delete for main rollout)
16
+
17
+ Q1: Why does this model exist in `gg-hf`?
18
+
19
+ A1: See https://docs.google.com/document/d/1bKaOw2mJPJDYhgFQGGVOyBB3M4Bm_Q3PMrfQeqeYi0M (Google internal only).
20
+
21
+ Q2: What does "SAE" mean?
22
+
23
+ A2: Sparse Autoencoder. See https://docs.google.com/document/d/1roMgCPMPEQgaNbCu15CGo966xRLToulCBQUVKVGvcfM (should be available to trusted HuggingFace collaborators, and Google too).
24
+
25
+ TODO(conmy): remove this when making the main repo.
26
+
27
+ ## 4. Point of Contact
28
+
29
+ Point of contact: Arthur Conmy
30
+
31
+ Contact by email:
32
+
33
+ ```python
34
+ ''.join(list('moc.elgoog@ymnoc')[::-1])
35
+ ```
36
+
37
+ HuggingFace account:
38
+ https://huggingface.co/ArthurConmyGDM