INDEX

Explanations

to music and sounds

The neuron fires on numeric tokens—especially measurements, counts, times, and other number-like values (including decimals) in the text.

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

’

-1.69

byshire

-1.43

 outras

-1.37

but

-1.35

 ''){

-1.32

you

-1.30

 niebla

-1.30

찐

-1.29

 você

-1.28

しますね

-1.27

POSITIVE LOGITS

 Mittelpunkt

1.58

 will

1.52

YYYY

1.52

はもちろん

1.45



1.43

澌

1.40

 ducato

1.38

 nuov

1.37

atecas

1.35

eavor

1.35

Activations Density 0.015%