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

 desper

-0.75

 mosqu

-0.74

 paran

-0.73

adesh

-0.71

acqu

-0.71

åĳ

-0.70

agy

-0.69

 occas

-0.68

 Scorp

-0.65

 Bagg

-0.65

POSITIVE LOGITS

ribute

0.72

ernels

0.67

laim

0.65

pec

0.62

uine

0.62

conom

0.61

 disproportionately

0.61

lations

0.61

hetical

0.61

ym

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.