INDEX

Explanations

html structure

This neuron responds to markup and code‐related tokens—especially XML/HTML tags, configuration keywords, and other programming-style identifiers.

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

ོ་

0.59

 বন

0.57

 descargar

0.55

 clearInterval

0.54

 оставить

0.52

noduch

0.52

 IconButton

0.52

 ماہر

0.52

ྦ

0.52

ിയും

0.51

POSITIVE LOGITS

 <!--

1.53

<!--

1.48

1.28

 <!--<

1.27

</

1.22

<!

1.15

<!

1.11

</

1.01

-<

1.00

Activations Density 0.002%