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

Synopsis

-0.84

olean

-0.84

 tremend

-0.73

ONSORED

-0.71

 cumbers

-0.71

nyder

-0.66

False

-0.65

ulously

-0.64

umbledore

-0.64

anas

-0.63

POSITIVE LOGITS

bies

0.75

tein

0.71

Ã°

0.70

ersen

0.65

armac

0.65

apers

0.64

 Tsuk

0.63

Street

0.63

 Astral

0.62

Aki

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.