INDEX

Explanations

the of

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ten

-0.98

 eight

-0.90

 nine

-0.89

 eleven

-0.82

 twelve

-0.81

 seven

-0.79

 dozen

-0.77

 zehn

-0.77

six

-0.74

 diez

-0.73

POSITIVE LOGITS

AsUp

0.67

pushd

0.57

iredo

0.57

0.56

forName

0.54

ContextHolder

0.54

ukone

0.52

Datuak

0.52

Selanjutnya

0.50

Representative

0.49

Activations Density 0.007%