INDEX

Explanations

code keywords and filenames

This neuron detects uppercase identifiers or constant‐style names (e.g. CSS variables, build settings, macros, enum-style tokens) in code and config snippets.

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

 okolí

-1.75

Тарих

-1.70

や

-1.69

 этого

-1.58

(!__

-1.54

鹛

-1.54

 této

-1.53

\}$.

-1.53

 vorigen

-1.52

 also

-1.50

POSITIVE LOGITS

”،

1.93

ifikasi

1.62

<bos>

1.59

——–

1.59

1.54

———-

1.48

 québéco

1.41

 souha

1.37

unpublished

1.34

◆◆

1.34

Activations Density 0.110%