INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.0.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Reb

-0.70

 Scotch

-0.63

 Fiji

-0.60

 Grind

-0.59

asketball

-0.59

 Barnett

-0.57

 Mole

-0.56

inity

-0.56

 recognised

-0.56

ITE

-0.55

POSITIVE LOGITS

éĹĺ

0.96

å£«

0.75

hetically

0.70

oyer

0.69

aeus

0.69

"},

0.69

fal

0.68

 appendix

0.67

catentry

0.67

ghazi

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.