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

MJ

-0.76

ighters

-0.73

aden

-0.72

IB

-0.72

ASE

-0.71

uction

-0.70

æĥ

-0.70

AMY

-0.70

iners

-0.69

ounced

-0.68

POSITIVE LOGITS

Merit

1.13

ĸļ

0.75

 Zucker

0.69

 Kass

0.68

ĪĴ

0.67

 Fork

0.66

bish

0.66

 Gunn

0.64

borg

0.63

 Gord

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.