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

CPC

-0.78

 defic

-0.78

Uncommon

-0.71

 Metropolitan

-0.69

 disbel

-0.68

æµ

-0.67

 Erie

-0.63

govtrack

-0.62

Ö¼

-0.62

 evils

-0.62

POSITIVE LOGITS

meter

0.70

imov

0.70

abi

0.68

tail

0.68

hemer

0.67

 Fired

0.65

anyon

0.64

ucket

0.64

nell

0.64

itage

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.