INDEX

Explanations

fun followed by noun

This neuron activates on blog‐style or promotional section headings—especially listicle titles (e.g. “7 Fun Ways…,” “Fun Fall Wreath DIY!”) tagged by prominent, title-case words.

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

 continúas

-1.60

any

-1.59

 ſch

-1.57

-1.50

哂

-1.48

the

-1.45

 dovr

-1.43

-1.42

 numerosi

-1.37

 ряда

-1.36

POSITIVE LOGITS

ErrUnexpectedEOF

1.59

":

1.55

 They

1.51

("

1.47

1.46

 raiſ

1.43

-"

1.40

 whack

1.38

粂

1.35

Fun

1.31

Activations Density 0.022%