INDEX

Explanations

months and publication details

This neuron flags numeric tokens—especially decimal numbers and numbers used in bibliographic metadata (volumes, issues, page numbers, dates, DOIs).

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

if

-1.42

/***

-1.26

ném

-1.25

७

-1.23

他

-1.23

しかない

-1.22

Caprio

-1.22

 other

-1.20

ρίζ

-1.18

armi

-1.17

POSITIVE LOGITS

 this

1.43

of

1.36

it

1.26

 which

1.24

our

1.23

 these

1.23

 that

1.14

 چقدر

1.13

–

1.13

鑰

1.07

Activations Density 0.060%