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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 stake

-0.78

eming

-0.78

seekers

-0.70

 awaits

-0.68

 Romero

-0.67

 apiece

-0.66

lette

-0.66

iton

-0.66

yond

-0.64

mire

-0.64

POSITIVE LOGITS

--------------------------------------------------------

0.76

@@@@

0.72

Cooldown

0.71

EStream

0.71

+++

0.70

ãĥīãĥ©

0.67

////////////////

0.67

++++++++

0.67

 Trace

0.66

artifacts

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.