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

uch

-0.76

olin

-0.75

ativity

-0.73

ule

-0.71

ilver

-0.71

stan

-0.69

culosis

-0.68

achu

-0.67

dden

-0.67

andise

-0.67

POSITIVE LOGITS

 Rodham

0.68

 patio

0.66

scape

0.65

 Playstation

0.63

 psychiat

0.61

 Weaver

0.60

bou

0.60

SourceFile

0.59

 cylinders

0.58

 dens

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.