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

 Metatron

-0.93

behind

-0.78

 dstg

-0.71

nee

-0.68

================================

-0.63

èĢħ

-0.61

Hidden

-0.61

advertisement

-0.60

 Millenn

-0.59

dain

-0.59

POSITIVE LOGITS

uary

0.74

emate

0.68

 Chambers

0.65

ovo

0.63

terness

0.63

eca

0.62

acle

0.62

atory

0.61

 Chamberlain

0.61

Sem

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.