© 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
PYTHIA-70M-DEDUPED · 4-RES-SM · 14628 ｜ Neuronpedia
Home
Under Peer Review · Sparse Autoencoders for Pythia-70M-Deduped
Pythia-70M-Deduped
Residuals
4-RES-SM
14628
Prev
Next
MODEL
4-res-sm
Source/SAE
INDEX
Go
Explanations
verbs and actions related to changing, creating, or managing settings and data
oai_token-act-pair · gpt-4o-mini
Triggered by @bot
No Scores
New Auto-Interp
AutoInterp Type
claude-4-5-haiku
Generate
Interpretable Score
Intrusion Detection
Top Features by Cosine Similarity
Configuration
ctigges/pythia-70m-deduped__res-sm_processed/4-res-sm
How To Load
Prompts (Dashboard)
32,768 prompts, 128 tokens each
Dataset (Dashboard)
monology/pile-uncopyrighted
Features
32,768
Data Type
torch.float32
Hook Name
blocks.4.hook_resid_post
Hook Layer
4
Architecture
standard
Context Size
128
Dataset
EleutherAI/the_pile_deduplicated
Activation Function
relu
Show All
Embeds
Show Plots
Show Explanation
Show Activations
Show Test Field
Show Steer
Show Link
IFrame
<iframe src="https://www.neuronpedia.org/pythia-70m-deduped/4-res-sm/14628?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/pythia-70m-deduped/4-res-sm/14628?embed=true&embedexplanation=true&embedplots=true&embedsteer=true&embedactivations=true&embedlink=true&embedtest=true
Not in Any Lists
Add to List
▼
No Comments
ADD
Neuron Alignment
Index
Value
% of L₁
307
+0.36
2.1%
156
+0.17
1.0%
306
+0.16
0.9%
Correlated Neurons
Index
P. Corr.
Cos Sim.
44
+0.36
0.04
307
+0.17
0.05
457
+0.16
0.03
Negative Logits
stown
-1.68
respectively
-1.58
ethic
-1.52
.”).
-1.49
NESS
-1.48
ruptcy
-1.46
.);
-1.46
.(
-1.41
.
-1.40
.”
-1.39
POSITIVE LOGITS
Ļª
4.21
ķ
3.65
µ
3.63
Ĵ
3.56
ĺ
3.54
IJ
3.50
ĨĴ
3.49
±
3.39
©
3.31
Ŀ
3.30
Act
ivations
Density 0.196%
Stacked
Snippet
Full
Show Breaks
Hide Breaks
No Known Activations
