INDEX

Explanations

even followed by verbs or the

The neuron detects emphasis or limiting adverbs—especially tokens like “even,” “ever,” and similar intensifiers.

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

-1.74

to

-1.72

of

-1.54

 และ

-1.45

-1.41

 Nosotros

-1.40

 toutes

-1.39

 REALLY

-1.38

WHY

-1.38

 остальных

-1.38

POSITIVE LOGITS

charpe

1.53

 Pamphlet

1.49

 kwamen

1.46

Manfaat

1.45

兩種

1.42

➅

1.41

ığını

1.41

autique

1.41

 techniczne

1.38

ècie

1.38

Activations Density 0.014%