INDEX

Explanations

but

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 zwar

-0.36

èĻ½çĦ¶

-0.33

åĽºçĦ¶

-0.31

èĻ½

-0.28

èĻ½è¯´

-0.28

åı¯æĺ¯

-0.26

 therefore

-0.25

æĢ§ä»·

-0.25

 though

-0.23

à¹ģà¸ķ

-0.23

POSITIVE LOGITS

 nevertheless

0.42

chers

0.41

 nonetheless

0.40

chedulers

0.34

åĲĮæĹ¶ä¹Ł

0.32

lers

0.28

ler

0.28

åį´æĺ¯

0.27

 toch

0.26

declaring

0.25

Activations Density 1.360%