INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

IER

-0.80

ROM

-0.68

 Ambrose

-0.63

 Franch

-0.62

Beast

-0.61

Klu

-0.59

 eBook

-0.59

ghazi

-0.58

 Operating

-0.58

sap

-0.58

POSITIVE LOGITS

illance

0.84

maxwell

0.80

NetMessage

0.78

::::

0.74

Interstitial

0.74

clerosis

0.73

ilts

0.73

aughter

0.72

izabeth

0.70

ourced

0.69

Activations Density 0.000%

No Known Activations

This feature has no known activations.