INDEX

Explanations

latex declarations and definitions

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

if

-2.20

 before

-1.85

 after

-1.83

 provide

-1.72

 have

-1.70

 when

-1.66

 create

-1.63

！（

-1.63

 begin

-1.62

any

-1.60

POSITIVE LOGITS

 marvelous

1.78

 unbelievably

1.74

 exceptionally

1.73

すっ

1.65

 wonderfully

1.65

 astonishing

1.63

 incredibly

1.63

 amazingly

1.62

 delightfully

1.60

 strikingly

1.57

Activations Density 0.001%