INDEX

Explanations

performance issues

word usage and structure

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

 concl

0.43

 versus

0.43

 Mughal

0.43

 Ajay

0.42

」。

0.42

 multin

0.42

 factorial

0.41

 edifice

0.41

 meridian

0.41

 преду

0.40

POSITIVE LOGITS

सी

0.60

 गरीबी

0.48

各地

0.46

Duke

0.45

LAN

0.43

 अस्पताल

0.42

Bristol

0.42

ീസ്

0.42

ले

0.41

En

0.41

Activations Density 0.001%