INDEX

Explanations

code imports

The neuron fires on code‐style identifiers and keywords—tokens like class names, function/variable names, import paths, etc.

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

or

-2.25

in

-2.06

to

-1.77

don

-1.66

as

-1.53

not

-1.51

="

-1.50

be

-1.48

-1.43

 without

-1.38

POSITIVE LOGITS

 fuis

1.62

 abiye

1.59

)();

1.47

bueno

1.46

 capilar

1.45

 troupes

1.43

 Tél

1.41

 contextLoads

1.41

 filtres

1.40

 FREDERICK

1.39

Activations Density 0.044%