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

Disk

-0.70

Dim

-0.68

anooga

-0.64

 bench

-0.62

Legend

-0.60

 Faith

-0.59

areth

-0.58

 cloth

-0.58

 Chips

-0.57

Score

-0.56

POSITIVE LOGITS

xual

0.92

ASA

0.82

 Baird

0.73

ãĥĢ

0.73

ilib

0.71

assetsadobe

0.71

teness

0.70

Ñģ

0.67

xon

0.67

vich

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.