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

 فريبيس

-0.73

)_/¯

-0.66

 Drapeau

-0.64

 nakalista

-0.63

WebServlet

-0.63

RegressionTest

-0.62

曖昧さ回避

-0.60

 يتيمه

-0.60

GOTREF

-0.59

NameInMap

-0.59

POSITIVE LOGITS

 latest

0.84

latest

0.75

 recent

0.70

recent

0.66

 labor

0.65

Latest

0.64

 presidency

0.64

labour

0.63

Recent

0.58

 those

0.58

Activations Density 0.004%