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

 Cars

-0.68

 motorists

-0.67

mine

-0.66

 Enrique

-0.65

rm

-0.65

bly

-0.63

affected

-0.62

him

-0.62

dam

-0.61

heon

-0.60

POSITIVE LOGITS

natureconservancy

0.81

Magazine

0.77

Accessory

0.76

 masc

0.74

DragonMagazine

0.73

enhagen

0.69

 magnification

0.69

Magikarp

0.69

Introduced

0.69

utenberg

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.