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

irable

-0.81

 Dise

-0.76

orthy

-0.73

schild

-0.70

arde

-0.67

Âł

-0.66

 Âł Âł

-0.65

etsk

-0.64

enance

-0.63

erd

-0.63

POSITIVE LOGITS

Gender

0.76

osate

0.76

hander

0.71

ieth

0.68

utsche

0.66

Reply

0.65

BuyableInstoreAndOnline

0.65

Balt

0.64

Forward

0.63

GAME

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.