INDEX

Explanations

license and notice information

The neuron activates on tokens that serve as field names or section‐heading labels (e.g. “type,” “price,” “license,” “notice,” “content‐type,” “conditions”) in structured or metadata‐rich text.

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

 cocinar

-0.93

or

-0.90

 doba

-0.90

ucine

-0.88

dollis

-0.86

 Vermögen

-0.85

🛕

-0.83

 dangers

-0.83

 jsem

-0.83

 cuidad

-0.81

POSITIVE LOGITS

of

1.68

is

1.15

 của

0.99

you

0.85

our

0.82

 noastre

0.82

的就是

0.82

 nostre

0.81

glLoad

0.81

 يلي

0.79

Activations Density 0.000%