INDEX

Explanations

lists and contrasts

The neuron fires on numeric tokens, especially quantities and measurements (integers or decimals) in the text.

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

 också

-1.53

 Bretaña

-1.52

Выводы

-1.45

壠

-1.43

Транспорт

-1.40



-1.40

 صلى

-1.39

Connecting

-1.35

㌧

-1.33

Заключение

-1.32

POSITIVE LOGITS

 just

2.08

 only

1.95

 then

1.87

 here

1.55

but

1.52

 same

1.45

one

1.44

 simply

1.40

 those

1.38

 femminile

1.33

Activations Density 0.109%