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

versions

-0.73

Newsletter

-0.67

cano

-0.66

 neurot

-0.65

ulsion

-0.65

cca

-0.64

 VIDEOS

-0.64

 Franch

-0.63

canon

-0.63

acas

-0.63

POSITIVE LOGITS

0.80

tie

0.71

bar

0.71

eday

0.68

rim

0.67

imony

0.67

come

0.67

lain

0.66

iciary

0.66

marked

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.