Neuronpedia
Get Started
API
Releases
Jump To
Search
Models
Circuit Tracer
NEW
Steer
SAE Evals
Exports
Blog
Slack
Privacy & Terms
Contact
Sign In
© Neuronpedia 2025
Privacy & Terms
Blog
GitHub
Slack
Twitter
Contact
Home
Joseph Bloom · Feature Splitting for GPT2-Small
GPT2-Small
Residual Feature Splitting 3072
8-RES_FS3072-JB
304
Prev
Next
MODEL
8-res_fs3072-jb
Source/SAE
INDEX
Go
Explanations
questions starting with "How does" or "Does."
oai_token-act-pair · gpt-3.5-turbo
Triggered by @bot
No Scores
New Auto-Interp
AutoInterp Type
claude-3-5-haiku-20241022
Generate
Top Features by Cosine Similarity
Comparing With
GPT2-SMALL @ 8-res_fs3072-jb
Configuration
jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072
How To Load
Prompts (Dashboard)
24,576 prompts, 128 tokens each
Dataset (Dashboard)
Skylion007/openwebtext
Features
3,072
Data Type
torch.float32
Hook Point
blocks.8.hook_resid_pre
Architecture
standard
Context Size
128
Dataset
Skylion007/openwebtext
Hook Point Layer
8
Activation Function
relu
Show All
Embeds
Plots
Explanation
Show Test Field
Default Test Text
IFrame
<iframe src=https://www.neuronpedia.org/gpt2-small/8-res_fs3072-jb/304?embed=true&embedexplanation=true&embedplots=true&embedtest=true" title="Neuronpedia" style="height: 300px; width: 540px;"></iframe>
Link
https://www.neuronpedia.org/gpt2-small/8-res_fs3072-jb/304?embed=true&embedexplanation=true&embedplots=true&embedtest=true
Not in Any Lists
Add to List
▼
No Comments
ADD
Neuron Alignment
Index
Value
% of L₁
311
+0.14
0.6%
493
+0.10
0.4%
88
+0.09
0.4%
Correlated Neurons
Index
P. Corr.
Cos Sim.
449
+0.14
0.07
287
+0.10
0.06
60
+0.09
0.06
Negative Logits
hig
-0.73
bis
-0.68
xon
-0.68
iken
-0.68
arer
-0.65
bsp
-0.64
ullivan
-0.64
lla
-0.62
iem
-0.61
psc
-0.61
POSITIVE LOGITS
?!
1.15
?
1.13
?]
1.12
?),
1.11
?)
1.10
?!"
1.08
?"
1.04
?:
1.02
?).
1.02
?",
1.01
Act
ivations
Density 0.421%
Test
Steer
Stacked
Snippet
Full
Show Breaks
Hide Breaks
No Known Activations
