INDEX

Explanations

contact details

The neuron flags contact‐info cues—headings and tokens like “Office,” “Email,” “Fax,” “Tel,” “@,” etc., that occur in address/phone/email blocks.

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

providedIn

-1.51

styleUrls

-1.45

oglobin

-1.38

]})

-1.32

 }));

-1.32

 tinha

-1.32

}],

-1.32

aning

-1.31

 éton

-1.31

 brillantes

-1.29

POSITIVE LOGITS

if

1.65

 also

1.53

 επίσης

1.52

 there

1.42

 również

1.41

 while

1.38

 aswell

1.30

1.28

 until

1.27

adalu

1.23

Activations Density 0.045%