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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rill

-0.81

itu

-0.76

 Nanto

-0.74

iants

-0.73

ilage

-0.72

aker

-0.70

agon

-0.70

glas

-0.69

 Pacers

-0.68

anta

-0.68

POSITIVE LOGITS

soDeliveryDate

0.70

²¾

0.66

 confinement

0.65

lihood

0.64

OLOGY

0.64

english

0.63

 guiActiveUn

0.63

WARE

0.62

Pixel

0.62

 ancest

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.