INDEX

Explanations

very quick / initially

This neuron fires on procedural or quantitative discourse markers—words that introduce steps, conditions, frequencies, or quantities (e.g. “first,” “then,” “every,” “if,” “only,” numeric tokens).

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

–

-1.45

-1.40

“

-1.35

 essentially

-1.34

 мають

-1.34

-1.33

 швидко

-1.33

 numerous

-1.33

 various

-1.30

 enormous

-1.30

POSITIVE LOGITS

"",

1.69

ztes

1.67

Privatpatient

1.58

selben

1.58

鮠

1.54

chesse

1.52

饜

1.52

БИ

1.51

 vrouwen

1.50

 münchen

1.50

Activations Density 0.121%