© Neuronpedia 2026
Privacy & Terms
Blog
GitHub
Slack
Twitter
Contact
Neuronpedia
Jacobian Lens
NEW
Natural Language
Autoencoders
NEW
Assistant Axis
NEW
Circuit Tracer
UPDATE
Releases
Jump To
Models
Search
Steer
SAE Evals
Exports
Guides
API
Community
Blog
Privacy & Terms
Contact
Sign In
Home
Models
Gemma-3-27B-IT
gemma-3-27b-it
Google Deepmind
Jump to Source/SAE
16-gemmascope-2-res-16k
Source/SAE
Go
Jump to Feature
16-gemmascope-2-res-16k
Source/SAE
INDEX
Go
Releases
Gemma Scope 2: Suite of SAEs and Transcoders for Gemma 3
December 2025
Google DeepMind
gemma-scope-2
Attention Visualizer
HeadVis
(Luger, Kamath et al.)
Find Head By Metric
Metric & Number of Heads
Induction Score
Prev Token Score
Attention Entropy
Self Attention
Top 8
Click head to select
Layer 5
Head 16
Layer 8
Head 23
Layer 11
Head 0
Layer 17
Head 27
Layer 23
Head 18
Head 19
Layer 35
Head 6
Layer 59
Head 25
Select Head Manually
Layer
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
Head Index
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Circuit Tracing
?
Choose a Featured Graph to Open
Generate New Graph
Search Explanations
All
By Release
By Model
By Sources
MODEL
Show Dashboards
Hide Dashboards
Search via Inference
?
Inference is not currently not enabled for this model.
Request
Browse
MODEL
Residual Stream - 16k
LAYER
Features in
GEMMA-3-27B-IT
@
16-gemmascope-2-res-16k
Hover over a feature on the left to preview its details.
Click a feature to lock it and interact with it.
