INDEX

Explanations

or

The neuron fires on domain‐specific technical terms—especially programming and system‐related jargon like “storyboard,” “code,” “processing,” “system,” 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

糨

-1.13

 delightfully

-1.13

 ۱۰

-1.10

 aktivi

-1.09

itamos

-1.08

 marqués

-1.07

 полноцен

-1.06

 praktik

-1.03

 vivent

-1.03

۹

-1.03

POSITIVE LOGITS

or

3.75

 или

1.95

 either

1.78

 یا

1.73

 hoặc

1.62

或

1.61

 либо

1.59

したり

1.50

หรือ

1.49

ή

1.48

Activations Density 0.114%