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

 smoother

-0.76

arnaev

-0.68

Downloadha

-0.67

uld

-0.67

pload

-0.64

ħĭ

-0.63

ô

-0.63

tyard

-0.63

 underwater

-0.62

 Disapp

-0.62

POSITIVE LOGITS

found

0.76

esa

0.72

uras

0.70

bra

0.68

caps

0.68

][

0.66

idas

0.66

Hug

0.66

omas

0.65

camp

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.