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

Enlarge

-0.71

 corrid

-0.68

isms

-0.67

 shenan

-0.67

xual

-0.64

DragonMagazine

-0.64

 ðŁĻĤ

-0.63

!:

-0.62

issance

-0.62

 behav

-0.61

POSITIVE LOGITS

 their

1.48

1.34

its

1.24

the

1.21

an

1.17

our

1.16

 this

1.11

his

1.08

 theirs

1.00

 another

1.00

Activations Density 0.000%

No Known Activations

This feature has no known activations.