INDEX

Explanations

numbers or punctuation

finding or trying to get information

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 Talk

0.49

on

0.48

gel

0.46

 Writing

0.46

 Sounds

0.45

 seaside

0.45

 Table

0.44

 Ball

0.44

 People

0.44

to

0.43

POSITIVE LOGITS

ienced

0.48

 DEPENDENCIA

0.46

 иной

0.46

 мир

0.45

 význam

0.45

 зміни

0.44

 друго

0.43

ర

0.43

мию

0.43

 สํา

0.42

Activations Density 0.004%