INDEX

Explanations

is a descriptor

The neuron reliably activates on the copula “is,” especially where it introduces a definition or descriptive statement.

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

はこんな感じ

-1.70

Ketua

-1.63

)|^{

-1.55

 semblent

-1.53

 зді

-1.51

Tout

-1.50

Mereka

-1.49

Sebelum

-1.48

Setelah

-1.45

Миха

-1.41

POSITIVE LOGITS

you

1.75

 your

1.62

1.56

1.49

 usein

1.47

rativa

1.38

the

1.37

魨

1.30

 that

1.29

 websites

1.28

Activations Density 0.112%