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

 Breed

-0.65

fifth

-0.64

enz

-0.63

ared

-0.63

cca

-0.62

rower

-0.62

 moth

-0.61

yne

-0.61

aer

-0.60

 emphasis

-0.60

POSITIVE LOGITS

Ĥİ

0.72

berra

0.71

 Wellington

0.66

 proble

0.66

Tickets

0.66

¶ħ

0.65

\\\\\\\\

0.64

sonian

0.64

"]=>

0.64

Īè

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.