INDEX

Explanations

most likely popular

political or positional phrases

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 нового

0.51

新たに

0.51

 새로운

0.50

 חד

0.48

嶄

0.47

leaved

0.47

新た

0.47

chedules

0.47

ă

0.47

ŏ

0.47

POSITIVE LOGITS

 universal

0.47

 understand

0.46

 сожалению

0.45

 water

0.41

 convain

0.41

 research

0.40

 diagnosis

0.40

Sol

0.39

 Mason

0.39

:-)

0.39

Activations Density 0.001%