INDEX

Explanations

hacking and breaking into systems

AI generating disinformation or tailored content

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

関

0.56

Fü

0.52

 Traditions

0.50

 Negeri

0.49

 Taban

0.48

 Specialty

0.48

Daw

0.48

 Furniture

0.48

 Terima

0.47

薪

0.47

POSITIVE LOGITS

номет

0.51

0.49

 axially

0.46

 antagonists

0.44

 piqu

0.44

 scratches

0.44

0.43

 scratched

0.42

rical

0.42

concatenate

0.41

Activations Density 0.000%