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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

WWF

-0.78

 cartoons

-0.72

 Kuala

-0.68

 Plain

-0.65

ONDON

-0.64

 Budapest

-0.62

 Ohio

-0.62

bred

-0.61

HIS

-0.61

 Disneyland

-0.61

POSITIVE LOGITS

lease

0.87

lyak

0.80

ursive

0.73

ascript

0.69

spin

0.69

agy

0.68

omo

0.67

ipt

0.66

utations

0.65

apt

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.