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

 liner

-0.76

pecially

-0.74

-0.71

 contamin

-0.65

retty

-0.65

 certific

-0.65

 monop

-0.63

ishly

-0.63

orously

-0.62

 param

-0.60

POSITIVE LOGITS

 Fathers

0.74

ador

0.73

utral

0.73

tics

0.72

ãĥ¼ãĥ³

0.72

 Deliver

0.67

 Workers

0.67

xon

0.66

zek

0.66

 Employees

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.