INDEX

Explanations

potential and possibility

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

 veckan

-1.24

有一

-1.18

or

-1.16

거나

-1.14

一般

-1.13

 estekak

-1.10

好多

-1.08

専門

-1.06

 Indep

-1.05

bbero

-1.05

POSITIVE LOGITS

 surpre

1.51

 custa

1.48

1.46

³)

1.39

 ajud

1.33

 ouvido

1.29

 estranho

1.27

ruary

1.26

 outside

1.25

uckles

1.25

Activations Density 0.086%