© Neuronpedia 2026
Privacy & Terms
Blog
GitHub
Slack
Twitter
Contact
Neuronpedia
Natural Language
Autoencoders
NEW
Assistant Axis
NEW
Circuit Tracer
UPDATE
Releases
Jump To
Search
Models
Steer
SAE Evals
Exports
Guides
API
Community
Blog
Privacy & Terms
Contact
Sign In
GEMMA-2-27B · 10-GEMMASCOPE-RES-131K · 11073 ｜ Neuronpedia
Home
Gemma-2-27B
10-GEMMASCOPE-RES-131K
11073
Prev
Next
MODEL
10-gemmascope-res-131k
Source/SAE
INDEX
Go
Explanations
names and technical terms
np_acts-logits-general · gemini-2.5-flash-lite
No Scores
proper nouns, especially names of people and software/technical terms.
oai_token-act-pair · claude-4-5-sonnet
Triggered by @bfgcn7frf5
No Scores
The neuron activates on capitalized named entities and acronyms (proper nouns, technical abbreviations, or similarly uppercase tokens).
oai_token-act-pair · o4-mini
Triggered by @jyhe0408
No Scores
New Auto-Interp
AutoInterp Type
claude-4-5-haiku
Generate
Top Features by Cosine Similarity
Configuration
google/gemma-scope-27b-pt-res/layer_10/width_131k
Prompts (Dashboard)
24,576 prompts, 128 tokens each
Dataset (Dashboard)
monology/pile-uncopyrighted
No Configuration Found
Show All
Embeds
Show Plots
Show Explanation
Show Activations
Show Test Field
Show Steer
Show Link
IFrame
<iframe src="https://www.neuronpedia.org/gemma-2-27b/10-gemmascope-res-131k/11073?embed=true&embedexplanation=true&embedplots=true&embedsteer=true&embedactivations=true&embedlink=true&embedtest=true" title="Neuronpedia" style="height: 300px; width: 540px;"></iframe>
Link
https://www.neuronpedia.org/gemma-2-27b/10-gemmascope-res-131k/11073?embed=true&embedexplanation=true&embedplots=true&embedsteer=true&embedactivations=true&embedlink=true&embedtest=true
Not in Any Lists
Add to List
▼
No Comments
ADD
Negative Logits
寕
-2.64
ที่
-2.30
需要
-2.25
⃢
-2.22
𖣘
-2.20
ၜ
-2.19
grüne
-2.09
combos
-2.05
入って
-2.05
beliebt
-1.98
POSITIVE LOGITS
腚
2.44
ENT
2.36
desee
2.33
嚶
2.33
婓
2.27
zra
2.22
珰
2.20
mesta
2.17
como
2.17
ους
2.16
Act
ivations
Density 0.018%
Stacked
Snippet
Full
Show Raw Tokens
Show Formatted
Show Breaks
Hide Breaks
No Known Activations
