INDEX

Explanations

edges and borders

The neuron fires on nouns naming decorative architectural or building‐feature elements (e.g. molding, eaves, brackets, mantel).

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

鹩

-1.64

THERE

-1.57

いかがでしたか

-1.56

놋

-1.56

didn

-1.56

-1.52

 تحتوي

-1.51

킁

-1.51

ného

-1.47

 girasol

-1.47

POSITIVE LOGITS

by

1.77

to

1.53

权威

1.40

 чле

1.36

 which

1.36

 alimenti

1.35

千年

1.35

 meil

1.34

諳

1.33

±

1.31

Activations Density 0.012%