INDEX

Explanations

how or whether to proceed

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

を用いる

0.18

 auront

0.18

 والك

0.18

９

0.18

 करताना

0.17

 высокой

0.17

স্তক

0.17

 arrivent

0.17

 longues

0.16

 vorm

0.16

POSITIVE LOGITS

how

0.54

 what

0.48

 whether

0.44

how

0.43

what

0.42

why

0.41

 কিভাবে

0.38

hva

0.38

 cómo

0.38

 hvordan

0.37

Activations Density 0.756%