INDEX

Explanations

referred to

The neuron fires on formal bureaucratic or regulatory language—especially agency names, acronyms, and mandated reporting or compliance terms.

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

ׇ

-1.75

 блюдо

-1.39

 mesmas

-1.38

 their

-1.37

 even

-1.35

 گیاه

-1.33

に於

-1.33

 प्रोड

-1.32

at

-1.31

as

-1.26

POSITIVE LOGITS

ּוֹ

1.89

ִּ

1.69

англ

1.59

ְּ

1.55

 اندروید

1.54

 آیفون

1.45

וּ

1.44

ίη

1.42

 flexib

1.41

וֹ

1.41

Activations Density 0.240%