INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

heast

-0.68

umenthal

-0.67

 plateau

-0.66

 helic

-0.65

ophe

-0.64

 hone

-0.64

 nightmare

-0.63

apon

-0.62

 Schedule

-0.62

 guru

-0.62

POSITIVE LOGITS

 Constantin

0.75

 Valiant

0.73

Wem

0.70

 Erit

0.67

Austral

0.67

 Nasa

0.67

mere

0.66

oeuv

0.65

 Ambro

0.64

 Spock

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.