INDEX

Explanations

be followed by adjective or participle

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

of

-1.66

 from

-1.63

or

-1.61

 kutoka

-1.52

for

-1.50

who

-1.48

 ambao

-1.45

ghalaya

-1.44

it

-1.34

 lakini

-1.30

POSITIVE LOGITS

鬣

1.38

1.30

 miscellaneous

1.28

Honors

1.28

 pasaj

1.24

ра

1.23

 asunto

1.23

 когда

1.21

 даже

1.20

 أنها

1.20

Activations Density 0.151%