INDEX

Explanations

This neuron detects language used for official announcements or calls to action—terms like “important,” “announcements,” “inviting,” “sign up,” “volunteer,” and requests for names/email.

future announcements and decisions

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

recently

-1.16

 Halen

-0.96

earlier

-0.95

 COMPOUNDS

-0.95

 منذ

-0.95

СЬ

-0.89

 koni

-0.88

volved

-0.87

yce

-0.87

fect

-0.85

POSITIVE LOGITS

 future

1.59

new

1.49

 potential

1.27

 plans

1.23

今後の

1.15

ligheter

1.13

 prospective

1.12

pantalon

1.10

 anew

1.08

 promising

1.06

Activations Density 0.054%