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

 Afgh

-0.70

TEXTURE

-0.68

¥µ

-0.67

 Rohingya

-0.66

 holiest

-0.65

Peb

-0.63

EFF

-0.63

 Bulgarian

-0.62

AAP

-0.62

 clinically

-0.60

POSITIVE LOGITS

eries

0.85

heid

0.78

sburgh

0.76

eus

0.75

 Vander

0.75

imony

0.74

pread

0.74

auc

0.74

iod

0.73

alia

0.73

Activations Density 0.000%

No Known Activations

This feature has no known activations.