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

oil

-0.79

ained

-0.77

assed

-0.71

itcher

-0.70

 undert

-0.68

atch

-0.67

iago

-0.66

acer

-0.65

ikan

-0.65

akia

-0.65

POSITIVE LOGITS

uten

0.78

Wem

0.77

orem

0.71

 Refuge

0.67

 Burma

0.65

xual

0.62

Peb

0.61

Nex

0.61

BSD

0.59

Dyn

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.