INDEX

Explanations

words that express emotions or opinions

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 nakalista

-0.71

GraphicsUnit

-0.66

 noDo

-0.66

 nationaux

-0.63

 TAMBIÉN

-0.62

 estekak

-0.61

 propOrder

-0.60

 Administrativna

-0.59

Бахар

-0.59

 kasarigan

-0.59

POSITIVE LOGITS

tling

0.54

λεύ

0.51

underbrace

0.47

aroo

0.47

bcryptjs

0.47

“

0.46

"]).

0.44

ensch

0.44

PED

0.44

rictive

0.43

Activations Density 1.660%