INDEX

Explanations

academic proceedings and journal titles

The neuron is detecting bibliographic citation markers, i.e. references to journals or proceedings (abbreviated journal names, “Proc.,” “Trans.,” volume/page indicators).

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

 أو

-1.73

 corregir

-1.72

<i>

-1.67

 使用

-1.64

-1.59

 then

-1.55

全

-1.52

 evaluar

-1.52

</h1>

-1.49

 новую

-1.49

POSITIVE LOGITS

of

1.99

 from

1.68

boneca

1.60

樣子

1.52

 夜景

1.52

brinco

1.50

ון

1.48

鼕

1.45

AutoModerator

1.43

yecto

1.43

Activations Density 0.034%