articles preceding words related to beginnings or important events

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

:✨

-0.96

homonymie

-0.94

Personensuche

-0.88

 виправивши

-0.87

 igång

-0.84

tvguidetime

-0.82

 Vikipedi

-0.75

adpleegd

-0.75

 jScrollPane

-0.73

utilisons

-0.73

POSITIVE LOGITS

<bos>

0.72

0.52

0.48

 pass

0.47

 times

0.47

0.46

0.45

Activations Density 2.738%