INDEX

Explanations

words in multiple languages

geographical locations and code

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

texte

0.46

enic

0.44

등

0.43

학

0.43

formul

0.42

 renseignements

0.42

ಣ್ಣ

0.41

Poland

0.41

erat

0.41

 등을

0.41

POSITIVE LOGITS

 означает

0.51

 incurs

0.48

 गैर

0.48

क्योंकि

0.46

 implies

0.46

 muszą

0.46

 karena

0.45

豈

0.44

 BECAUSE

0.44

 означа

0.43

Activations Density 0.001%