INDEX

Explanations

robot forms and media titles

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

 however

-1.67

 però

-1.57

 incluindo

-1.23

讓我

-1.22

dây

-1.16

 fábricas

-1.15

麻煩

-1.14

hộp

-1.13

 الجهاز

-1.12

 אבל

-1.11

POSITIVE LOGITS

of

2.42

 '/';

1.55

 '-';

1.37

 engels

1.31

 kablo

1.22

ɵɵelement

1.19

kuh

1.17

 beina

1.16

 geograf

1.16

гие

1.16

Activations Density 0.027%