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

ring

-0.56

 gross

-0.52

itis

-0.52

 billing

-0.52

anga

-0.51

screen

-0.51

 targeting

-0.51

iah

-0.51

emon

-0.51

 fore

-0.50

POSITIVE LOGITS

you

1.53

you

1.21

YOU

1.15

YOU

0.96

ya

0.95

we

0.89

yip

0.85

You

0.80

 thee

0.74

You

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.