INDEX

Explanations

numerical codes and identifiers

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

the

-3.09

to

-2.66

 alcune

-2.25

 decisões

-2.03

 speichern

-2.00

 questões

-1.98

-1.95

୬

-1.93

 notwendig

-1.93

 várias

-1.92

POSITIVE LOGITS

 Give

1.96

یزات

1.87

他们

1.85

iniest

1.84

 их

1.82

);

1.82

1.80

Those

1.78

”،

1.77

and

1.76

Activations Density 0.177%