INDEX

Explanations

leading provider or LIABLE

The neuron fires on content‐bearing descriptive words—especially adjectives or qualifiers like “greater,” “pseudo,” “leading,” or “liable” that introduce or emphasize key concepts.

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

you

-1.73

 because

-1.66

 when

-1.63

the

-1.60

 accordance

-1.59

by

-1.52

 fellow

-1.39

 companyId

-1.34

 several

-1.34

 both

-1.33

POSITIVE LOGITS

 Primera

1.55

Après

1.53

產品

1.43

Пі

1.40

也会

1.38

న్

1.35

 porcelaine

1.34

Игра

1.33

Sortie

1.31

🟪

1.31

Activations Density 0.041%