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

ewitness

-0.69

 hybrids

-0.66

hammer

-0.65

age

-0.64

Rel

-0.62

ury

-0.60

 Zucker

-0.60

 cent

-0.60

amorph

-0.59

et

-0.59

POSITIVE LOGITS

uggest

0.88

Ô

0.83

tnc

0.78

 newsp

0.77

BuyableInstoreAndOnline

0.76

yout

0.76

 withd

0.74

 Citiz

0.73

ailability

0.72

 conclud

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.