INDEX

Explanations

will/would followed by a verb

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

-1.52

 maar

-1.45

-1.32

<!--

-1.31

 qual

-1.27

 cathédrale

-1.25

 essayé

-1.24

“

-1.23

佉

-1.23

POSITIVE LOGITS

 also

1.70

 همچنین

1.55

ξεις

1.43

 també

1.41

 consape

1.39

 nahil

1.38

 thereby

1.35

 構造

1.35

 también

1.35

蹌

1.34

Activations Density 0.051%