INDEX

Explanations

<strong>0</strong> followed by 1

The neuron fires on standalone numeric tokens (individual digits) in structured citation metadata—e.g. volume, issue, and year numbers in parentheses.

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

dịch

-1.46

different

-1.43

following

-1.39

中です

-1.39

 verschillende

-1.35

を読

-1.34

をご覧

-1.34

 différentes

-1.34

を書く

-1.34

を読む

-1.33

POSITIVE LOGITS

),

1.48

</strong>

1.31

 undeniably

1.27

عارف

1.24

pasti

1.20

 brief

1.17

 gjelder

1.16

 from

1.16

 概要

1.16

żdż

1.15

Activations Density 0.104%