INDEX

Explanations

URLs and email subjects

dates formatted with commas separating day and year.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

</i>

-1.49

龔

-1.41

šanu

-1.38

tedì

-1.32

casila

-1.31

idorm

-1.31

 botanique

-1.30

pañol

-1.27

<bos>

-1.26

POSITIVE LOGITS

也许

1.59

</h4>

1.45

或许

1.42

 vitale

1.41

 отличаются

1.38

);

1.34

 problemet

1.34

 Saltar

1.27

 leeren

1.27

 grossa

1.26

Activations Density 0.007%