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

 Volunteers

-0.75

isu

-0.71

atha

-0.70

 Filip

-0.64

abulary

-0.64

 Maple

-0.61

ibly

-0.61

glers

-0.60

 Liberation

-0.59

iddles

-0.58

POSITIVE LOGITS

 horizont

0.72

merce

0.71

bey

0.66

DragonMagazine

0.65

 Seym

0.64

Chel

0.64

Wars

0.63

zyk

0.63

CAST

0.63

cha

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.