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

isconsin

-0.73

aph

-0.72

 mathemat

-0.70

renheit

-0.67

 vocational

-0.67

bilt

-0.65

 largeDownload

-0.65

 referen

-0.63

myra

-0.63

 creatine

-0.63

POSITIVE LOGITS

Media

1.20

wear

0.77

blocks

0.67

ench

0.67

 Preview

0.64

vironments

0.64

ener

0.62

mods

0.61

 Stun

0.61

guards

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.