INDEX

Explanations

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

 nahilalakip

-0.81

 real

-0.68

 tobacco

-0.68

 useAppContext

-0.68

 verbal

-0.67

 Ketten

-0.67

expandindo

-0.66

 Vanden

-0.66

ureau

-0.66

 dredging

-0.66

POSITIVE LOGITS

Cyfeiriadau

0.40

 Offisielt

0.39

 vänt

0.37

 acrylique

0.37

 bendera

0.36

 spettacolo

0.35

 gouttes

0.35

 chré

0.35

 gerichtet

0.35

 ongles

0.34

Activations Density 0.002%