INDEX

Explanations

links, foreign words, or specific jargon

The neuron is looking for URL‐ or link‐related tokens (e.g. URL parameter segments like “UU…” and link markers such as “external”).

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

and

-1.45

</b>

-1.45

qai

-1.41

 základní

-1.29

');

-1.27

 offizi

-1.27

 hogyan

-1.27

 folgenden

-1.27

 creatividad

-1.26

 suficientemente

-1.26

POSITIVE LOGITS

of

1.89

1.75

Bisa

1.69

 ventre

1.64

Awal

1.63

 rigide

1.50



1.48

cking

1.47

vertible

1.46

 Jumat

1.45

Activations Density 0.000%