INDEX

Explanations

numeric or alphanumeric codes

The neuron activates on backslash characters (often appearing in ASCII‐art, file paths, or similar backslash‐heavy sequences).

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

 experto

-1.46

 sowohl

-1.40

 llanto

-1.38

嚆

-1.38

 associée

-1.38

 DOMINGO

-1.34

 lección

-1.32

鹛

-1.31

劇情

-1.30

 mohl

-1.26

POSITIVE LOGITS

at

1.30

)]);

1.30

1.27

 дизе

1.26

</i>

1.24

acionais

1.23

in

1.23

</h1>

1.23

gebras

1.23

Indic

1.22

Activations Density 0.001%