INDEX

Explanations

myth slot, no surprise

names and technical terms

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

 Language

0.43

Ки

0.40

 храм

0.40

Centro

0.38

 tessuto

0.38

Source

0.37

Config

0.37

ewną

0.36

 लैंग्वेज

0.35

POSITIVE LOGITS

 borrowers

0.46

 swipes

0.45

 cannabis

0.44

 unapolog

0.41

 treason

0.40

 lenders

0.39

 prést

0.38

 litigants

0.38

 cannabinoids

0.38

 borrowings

0.37

Activations Density 0.001%