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

Ward

-0.70

dust

-0.69

belt

-0.68

Sa

-0.67

 Misty

-0.64

 Stew

-0.64

radio

-0.63

 Auction

-0.63

 Unity

-0.62

Wide

-0.62

POSITIVE LOGITS

per

1.27

 percent

1.23

 basis

1.07

 Percent

1.05

percent

0.86

 burden

0.84

 crore

0.79

 billion

0.78

%"

0.74

%.

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.