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

 Duck

-0.75

lee

-0.73

otos

-0.70

ãĥĺãĥ©

-0.70

 Vand

-0.69

atum

-0.68

ete

-0.68

 Vert

-0.67

 Niet

-0.67

iton

-0.66

POSITIVE LOGITS

 abundantly

0.80

quished

0.74

terms

0.73

 passions

0.68

 confir

0.67

ebin

0.67

CLOSE

0.67

 whirlwind

0.65

 recip

0.64

 grapp

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.