INDEX

Explanations

either express or implied

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ラップ

-0.78

 Kura

-0.73

髭

-0.73

ilov

-0.71

YCH

-0.69

topt

-0.68

 Trat

-0.66

アフ

-0.66

 Ganze

-0.65

Matchers

-0.64

POSITIVE LOGITS

bits

0.69

)--(

0.66

Roo

0.65

iegel

0.65

 strip

0.65

 bits

0.65

atisfaction

0.65

ophyllum

0.64

jící

0.62

Chancellor

0.62

Activations Density 0.091%