INDEX

Explanations

profile | address | tiktok

The neuron lights up on prominent topic words and proper-noun keywords—i.e. the main headings, branded or technical terms, and category labels you see in titles, menus, and product listings.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 apostila

0.66

 непосредственно

0.64

--*/

0.64

:["

0.63

 অধিনায়ক

0.63

PathResource

0.61

mergeddata

0.61

KeyboardButton

0.60

 automático

0.60

hoea

0.60

POSITIVE LOGITS

2.02

｜

1.59

\|

1.50

|,

1.39

|.

1.30

1.27

\|

1.18

1.17

::

1.12

$|

1.04

Activations Density 0.462%