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

nown

-0.82

acci

-0.79

merce

-0.75

ryce

-0.73

resa

-0.73

wcsstore

-0.70

artment

-0.69

ullivan

-0.69

rongh

-0.64

 chop

-0.64

POSITIVE LOGITS

ãĥİ

0.85

qt

0.75

pad

0.73

pt

0.72

plates

0.69

tex

0.68

pts

0.67

etus

0.66

 XIII

0.65

Zah

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.