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

DW

-0.68

 Bolshevik

-0.67

 Adventure

-0.67

 Portug

-0.65

elvet

-0.65

 Sark

-0.65

 Wave

-0.63

 Colonial

-0.63

 Raider

-0.63

 Ministry

-0.62

POSITIVE LOGITS

milo

0.78

eworthy

0.71

ingen

0.70

ancies

0.66

hern

0.66

Rating

0.65

 challeng

0.63

Includes

0.63

successfully

0.62

fired

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.