INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

vana

-0.73

dit

-0.73

lda

-0.65

azo

-0.65

pell

-0.65

 Neph

-0.65

idon

-0.65

cluding

-0.64

avy

-0.63

hammad

-0.62

POSITIVE LOGITS

 suspic

0.77

ÃĥÃĤ

0.75

 palp

0.72

 millenn

0.70

 earthqu

0.69

 conclud

0.67

cknow

0.66

dayName

0.66

 stret

0.65

acia

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.