INDEX

Explanations

structured object components

The neuron activates on numeric literals (numbers) in code or data structures.

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

providedIn

-1.88

 dicembre

-1.50

%).

-1.48

？

-1.43

by

-1.38

To

-1.38

ですが

-1.38

 rootReducer

-1.37

 незавер

-1.37

 second

-1.37

POSITIVE LOGITS

 anadolu

1.73

 türkiye

1.70

媃

1.63

鱚

1.55

分明

1.54

瞜

1.54

鼉

1.51

ǻ

1.50

 можешь

1.50

extendable

1.50

Activations Density 0.017%