INDEX

Explanations

descriptive phrases

The neuron fires on words that mark high-level document structure or metadata—e.g. section headings and labels like “Date,” “Includes,” “Core,” “Dedicated,” “Alternative,” 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

</h2>

-2.03

ówno

-1.38

 Stephanie

-1.38

ADDY

-1.35

ンピース

-1.35

"";

-1.34

leş

-1.30

ofre

-1.30

ocidade

-1.29

 Biografie

-1.26

POSITIVE LOGITS

 they

1.55

of

1.48

you

1.41

by

1.36

焢

1.26

 coö

1.23

 גם

1.22

作って

1.22

 Якщо

1.22

简约

1.21

Activations Density 0.292%