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

oad

-0.76

Child

-0.71

ails

-0.68

Ĥİ

-0.65

woman

-0.64

Assembly

-0.64

Cu

-0.63

Pool

-0.63

icans

-0.61

ouch

-0.61

POSITIVE LOGITS

 sovere

0.91

rolet

0.77

 territ

0.73

chart

0.67

 mosqu

0.65

Ðº

0.62

 violet

0.62

pole

0.61

SOM

0.61

ankind

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.