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

audi

-0.97

sembly

-0.78

anova

-0.73

wen

-0.72

borough

-0.71

omorphic

-0.71

inged

-0.70

urus

-0.68

haps

-0.66

rx

-0.65

POSITIVE LOGITS

 Burns

1.09

Initialized

0.70

ERY

0.66

Course

0.65

INE

0.63

 Cooking

0.61

ARP

0.60

OY

0.59

 Judicial

0.59

Appearance

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.