INDEX

Explanations

articles (a/an) followed by a word

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

<0xA3>

-1.30

所以在

-1.01

mixture

-1.01

aneous

-0.99

(/^

-0.98

 increase

-0.96

preneurs

-0.95

各式

-0.95

CCIÓN

-0.95

mix

-0.93

POSITIVE LOGITS

important

1.23

 century

1.22

 important

1.17

 otherwise

1.15

 столь

1.13

 already

1.13

otherwise

1.10

 something

1.09

 hitherto

1.09

 ormai

1.09

Activations Density 0.116%