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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 newsletters

-0.82

inelli

-0.80

illary

-0.73

Reviewed

-0.61

iott

-0.61

psc

-0.60

cou

-0.60

 Chest

-0.60

 mirror

-0.58

ItemThumbnailImage

-0.58

POSITIVE LOGITS

KEN

0.77

Ã°

0.75

perty

0.73

enium

0.71

dj

0.71

Ñĭ

0.67

sson

0.66

ãĤ¤

0.66

à¼

0.66

·

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.