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

 Typhoon

-0.73

Ago

-0.70

 Micro

-0.69

te

-0.67

lie

-0.66

ItemImage

-0.65

imaru

-0.65

 Gram

-0.64

agne

-0.64

uo

-0.63

POSITIVE LOGITS

oples

0.76

ASC

0.73

YP

0.71

rompt

0.69

 swast

0.68

Compat

0.68

 replay

0.65

 thrott

0.64

 positives

0.63

ivil

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.