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

 resil

-0.70

rophic

-0.69

chwitz

-0.67

aleb

-0.65

 flock

-0.65

estone

-0.65

)=(

-0.64

 fray

-0.63

rag

-0.63

 grav

-0.63

POSITIVE LOGITS

Phones

0.72

 permissions

0.71

LOCK

0.68

 Authentication

0.67

abet

0.67

BACK

0.66

phone

0.64

NPR

0.64

phones

0.63

LAN

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.