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

iol

-0.70

ieties

-0.70

iann

-0.69

 bargain

-0.68

icipated

-0.67

arters

-0.66

 Handling

-0.65

oulder

-0.65

iola

-0.65

iety

-0.65

POSITIVE LOGITS

 redd

0.79

 conclud

0.76

 distingu

0.70

 trave

0.67

fw

0.67

 dement

0.66

 sheds

0.66

 longest

0.65

had

0.65

aeda

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.