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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Terra

-0.66

 Seoul

-0.63

 Cannon

-0.62

 behalf

-0.62

stadt

-0.62

æ³

-0.61

soever

-0.60

each

-0.60

Pack

-0.59

Dest

-0.58

POSITIVE LOGITS

arthy

0.82

ptin

0.76

enhagen

0.71

laus

0.71

ensation

0.71

xual

0.70

insula

0.70

natureconservancy

0.68

hower

0.67

iosis

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.