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

 Neutral

-0.70

galitarian

-0.70

Õ

-0.69

Ï

-0.66

 Thief

-0.65

 Petty

-0.64

Î

-0.64

Îµ

-0.62

 IMAGES

-0.61

Ïī

-0.61

POSITIVE LOGITS

bats

0.81

andum

0.80

bowl

0.73

enhagen

0.72

 baseman

0.68

 tamp

0.68

packages

0.67

hiba

0.65

 kickoff

0.64

 sergeant

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.