INDEX

Explanations

to

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

be

-0.69

 kill

-0.68

 feed

-0.67

 assist

-0.66

 deal

-0.66

 starve

-0.66

 draw

-0.65

 exploit

-0.65

 change

-0.64

get

-0.63

POSITIVE LOGITS

 avoient

0.76

 étoit

0.68

 réguli

0.68

 marins

0.65

 confronti

0.63

 feroit

0.62

 igång

0.62

 complètes

0.62

 ouvertes

0.61

 démocr

0.60

Activations Density 0.002%