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

ech

-0.85

internet

-0.78

fields

-0.76

oral

-0.75

ospace

-0.73

iframe

-0.72

rose

-0.71

computer

-0.69

lÃ©

-0.68

comfort

-0.67

POSITIVE LOGITS

 Shiva

0.69

 Mend

0.67

 Adams

0.63

ND

0.63

 EDITION

0.63

 Fernand

0.62

 sidx

0.62

ãģ®ç

0.60

 Pist

0.60

 Sussex

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.