INDEX

Explanations

distances, diseases, historical places

The neuron activates on numeric measurements—especially distances, times, and other numerical metrics (e.g., “1000-meter,” “38.5,” “74,” “27.125”).

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

 When

-1.64

 there

-1.55

-1.52

 Fazit

-1.39

daten

-1.33

<h2>

-1.33

 only

-1.32

靦

-1.31

质疑

-1.31

 utilizando

-1.30

POSITIVE LOGITS

 vermitteln

1.63

lijke

1.60

cinogenicity

1.59

活力

1.55

”,

1.52

riebene

1.52

?"

1.51

these

1.50

will

1.49

 from

1.47

Activations Density 0.036%