INDEX

Explanations

any or anything

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

 also

-1.56

 certains

-1.37

 sometimes

-1.34

 некоторых

-1.30

 некоторые

-1.22

 potrebbero

-1.22

 VERY

-1.16

 ļoti

-1.15

 vissa

-1.15

 eventually

-1.12

POSITIVE LOGITS

any

3.61

 anything

3.23

有任何

2.95

任何

2.48

 quelconque

2.30

anything

2.23

 anywhere

2.20

 qualquer

2.16

 anyone

2.13

 herhangi

2.11

Activations Density 0.101%