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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Croat

-0.74

 largeDownload

-0.72

%:

-0.72

 Swift

-0.69

âĢ¢âĢ¢âĢ¢âĢ¢

-0.68

@#&

-0.67

DragonMagazine

-0.67

ANS

-0.66

 Citiz

-0.65

 Ukrainian

-0.64

POSITIVE LOGITS

ages

0.84

hig

0.83

hash

0.77

chi

0.76

agra

0.75

chio

0.73

aged

0.72

aging

0.68

 shaft

0.67

 Riders

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.