INDEX

Explanations

political statements and media manipulation

The neuron flags lines that serve as structural labels—titles, headings, or speaker/byline attributions often marked by quoted strings or colon‐terminated names.

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

ׇ

-3.39

 گیاه

-2.47

 mesmas

-2.39

 mães

-2.17

 RISERV

-2.14

 botanique

-2.09

 ویتامین

-2.08

 النبات

-2.05

 الجلد

-1.95

логические

-1.95

POSITIVE LOGITS

ִּ

2.02

ּוֹ

1.98

 оружи

1.91

 начин

1.88

 موبایل

1.74

ַּ

1.70

 médec

1.69

The

1.66

mathrm

1.66

וֹ

1.59

Activations Density 1.306%