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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Moons

-0.69

 Dino

-0.65

eno

-0.65

 Wrestling

-0.64

 Antar

-0.63

 passage

-0.62

 Reincarn

-0.62

 Raleigh

-0.61

 Trin

-0.60

 cabinets

-0.59

POSITIVE LOGITS

Interstitial

0.83

ilaterally

0.73

###

0.71

ï¸ı

0.69

ðĿ

0.68

ðŁ

0.66

sing

0.66

under

0.66

flo

0.65

sten

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.