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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Bezos

-0.70

African

-0.69

 Auburn

-0.69

 Pens

-0.66

 Azerbai

-0.65

Fu

-0.65

 Saddam

-0.63

 DeVos

-0.62

FSA

-0.61

 grandparents

-0.61

POSITIVE LOGITS

ethy

0.71

£ı

0.69

ForgeModLoader

0.68

 Rece

0.67

reet

0.67

rawler

0.66

½

0.66

stead

0.63

rint

0.63

auld

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.