© 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
Search
Steer
SAE Evals
Exports
interp-engine
NEW
Guides
API
Community
Blog
Privacy & Terms
Contact
Sign In
Home
Gemma-3-12B
24-GEMMASCOPE-2-RES-16K
6422
Prev
Next
MODEL
24-gemmascope-2-res-16k
Source/SAE
INDEX
Go
Explanations
would followed by verb/adjective
np_acts-logits-general · gemini-2.5-flash-lite
No Scores
The neuron strongly activates on the modal verb “would” (and similar high‐value instances like “never”)—i.e. it detects conditional or hypothetical statements.
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-2-12b-pt/resid_post/layer_24_width_16k_l0_medium
How To Load
Prompts (Dashboard)
392,802 prompts, 256 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-3-12b/24-gemmascope-2-res-16k/6422?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-3-12b/24-gemmascope-2-res-16k/6422?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
sẽ
1.14
ра
1.09
swürdig
1.00
の頃
1.00
о
0.98
Тому
0.96
тся
0.96
ECE
0.95
ИА
0.94
ρί
0.93
POSITIVE LOGITS
ج
1.28
ی
1.23
haya
1.08
ين
1.04
yuan
1.04
تى
1.02
ਾ
1.00
客様
0.98
یر
0.98
าค
0.97
Act
ivations
Density 0.088%
Test
Stacked
Snippet
Full
Show Breaks
Hide Breaks
No Known Activations
