INDEX

Explanations

This neuron fires on numeric and ordinal markers (digits and suffixes like “th”) typically found in formal document headings or section numbers.

question prefix or specific phrases

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

 блюдо

-1.52

 chré

-1.43

 بدأ

-1.34

 この

-1.29

֩

-1.26

inau

-1.22

 Besitzer

-1.21

gilet

-1.18

 Söz

-1.17

 roś

-1.16

POSITIVE LOGITS

ּוֹ

1.62

־ה

1.45

וּ

1.39

 trattano

1.34

Descriere

1.33

̳

1.30

enegger

1.29

ִּ

1.26

动静

1.25

нову

1.23

Activations Density 0.000%