INDEX

Explanations

inverter plays role

This neuron detects highly technical, process-descriptive language in engineering contexts—especially terms and verbs describing power-conversion components (e.g. “inverter,” “booster,” “convert,” “thereby,” “achieving,” “distributed”).

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

to

-2.17

 were

-1.78

 today

-1.67

primarily

-1.60

buon

-1.57

なども

-1.52

denver

-1.51

dubai

-1.50

plish

-1.49

carolina

-1.48

POSITIVE LOGITS

1.86

1.78

」、

1.77

1.59

uating

1.50

 мост

1.48

Ᏻ

1.48

1.47

:✨

1.46

Activations Density 0.371%