INDEX

Explanations

quality or level

comparable performance level

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

事

0.42

不思議

0.41

意外

0.40

чени

0.40

 oddly

0.40

 cautioned

0.39

 найбіль

0.39

boo

0.38

override

0.38

 অদ্ভুত

0.38

POSITIVE LOGITS

 caliber

1.97

 calibre

1.90

caliber

1.57

 수준

1.37

 niveau

1.34

 level

1.31

 مستوى

1.29

 stature

1.28

级别

1.24

 nível

1.23

Activations Density 0.141%