INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

netflix

-0.73

EVA

-0.73

SourceFile

-0.72

toe

-0.72

krit

-0.72

inventoryQuantity

-0.71

bryce

-0.70

iasis

-0.70

lag

-0.68

anuts

-0.68

POSITIVE LOGITS

alian

0.74

 Czech

0.66

 vocabulary

0.65

 fastest

0.64

 fluent

0.64

ocamp

0.63

 slender

0.61

 lips

0.61

estern

0.61

 advant

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.