INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ixties

-0.80

ties

-0.71

 actionGroup

-0.68

utenberg

-0.66

shed

-0.65

leans

-0.65

 footh

-0.64

ujah

-0.63

aret

-0.63

 Digest

-0.62

POSITIVE LOGITS

 Chamberlain

0.68

 Eisen

0.67

 Stat

0.63

 Calendar

0.63

heim

0.63

atars

0.63

 Wilde

0.62

ARM

0.61

APS

0.60

)!

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.