INDEX

Explanations

first book or series

The neuron detects ordinal or numeric mentions that specify an item’s position or count within a series (e.g., “first,” “two,” “five,” “second book,” “third in the series”).

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

 first

-1.75

 because

-1.59

 ....

-1.58

 only

-1.56

 both

-1.55

 .....

-1.55

 bereits

-1.49

by

-1.48

 expertly

-1.47

 very

-1.45

POSITIVE LOGITS

 Schritte

1.43

confused

1.40

 recentes

1.34

𐄁

1.31

 lokalen

1.31

 cerâmica

1.31

fibonacci

1.29

 artesanato

1.27

 Gespräche

1.27

как

1.27

Activations Density 0.041%