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

 synd

-0.76

 Tend

-0.72

pend

-0.71

DragonMagazine

-0.69

Quality

-0.68

 teasp

-0.67

Contents

-0.67

orkshire

-0.65

Rated

-0.65

Bone

-0.64

POSITIVE LOGITS

 manif

0.88

Ĥİ

0.72

opsis

0.70

uster

0.69

 manifold

0.69

»Ĵ

0.69

ntil

0.67

nels

0.66

 stimul

0.66

 elig

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.