INDEX

Explanations

conjunctions and list items

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

なんですよ

-1.46

Never

-1.32

なんですね

-1.31

 utaf

-1.30

 sababu

-1.28

 meninggal

-1.28

 respekt

-1.27

 goû

-1.27

 kommet

-1.26

discuss

-1.24

POSITIVE LOGITS

and

4.38

And

1.70

or

1.36

 और

1.21

и

1.21

1.15

\&

1.13

and

1.12

1.11

楽しめます

1.08

Activations Density 0.102%