INDEX

Explanations

congratulations on their achievements

The neuron fires on words and tokens used in congratulatory, celebratory announcements (e.g. “congrats,” “congratulations,” “on,” “your,” “first,” “victory,” etc.).

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

 both

-1.89

to

-1.82

 will

-1.80

 every

-1.77

 each

-1.68

 only

-1.66

up

-1.43

 throughout

-1.43

 very

-1.42

 various

-1.42

POSITIVE LOGITS

strich

1.51

Musique

1.43

 atelier

1.38

Pä

1.37

Télé

1.33

 eenige

1.33

 komik

1.29

 вигля

1.29

Jeux

1.28

ȍ

1.27

Activations Density 0.020%