INDEX

Explanations

assuming [variable] is positive

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 enorme

-1.66

裢

-1.57

voer

-1.55

 know

-1.53

尨

-1.51

am

-1.49

ſche

-1.48

不忘

-1.47

entious

-1.46

ung

-1.43

POSITIVE LOGITS

</h3>

1.74

 Genehmigung

1.59

 gewähr

1.53

 изменение

1.41

 Ironically

1.38

and

1.37

↵

1.36

หรือ

1.35

しましたが

1.34

or

1.33

Activations Density 0.014%