INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

ascript

-0.79

aths

-0.75

 expense

-0.72

ussia

-0.71

ovember

-0.70

iffe

-0.68

ongevity

-0.68

ook

-0.66

esse

-0.65

iors

-0.64

POSITIVE LOGITS

 numbered

0.76

²¾

0.73

worn

0.68

 morphed

0.67

 actionGroup

0.65

HOU

0.64

DragonMagazine

0.64

thouse

0.63

fur

0.62

 chanting

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.