INDEX

Explanations

examine

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

 view

-0.60

 examine

-0.56

 foncé

-0.55

 assess

-0.53

actéristique

-0.52

evaluate

-0.51

 fallu

-0.51

 evaluate

-0.50

ção

-0.49

view

-0.48

POSITIVE LOGITS

脚注の使い方

0.75

WebVitals

0.71

ValueStyle

0.70

 nakalista

0.68

UnusedPrivate

0.66

:✨

0.65

اریخ

0.63

 Magdal

0.59

Rüyada

0.57

FontStyle

0.57

Activations Density 0.151%