INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rha

-0.73

utic

-0.72

wagen

-0.68

thia

-0.66

acl

-0.63

 contrasted

-0.62

ulhu

-0.60

atum

-0.60

urat

-0.60

Afee

-0.60

POSITIVE LOGITS

 Cardinals

0.75

shire

0.73

sembly

0.63

 Louis

0.62

tein

0.62

holders

0.61

 Kenny

0.61

quez

0.60

 Blaze

0.59

enthal

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.