INDEX

Explanations

digital signal processing, distribution licenses, Quran types, quantification of, chaos theory, Berlin types, sweet Virgin

This neuron responds to high–value tokens—especially numerals, technical acronyms, and specialized jargon—typical of data values and scientific/engineering terminology.

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

 beiden

-1.86

Both

-1.86

两者

-1.70

 दोनों

-1.63

 entrambi

-1.57

 ambos

-1.56

 begge

-1.53

 Beide

-1.53

这两个

-1.52

どちらも

-1.51

POSITIVE LOGITS

all

2.08

etc

1.54

 three

1.30

 എന്ന

1.22

1.15

etc

1.15

三

1.10

～！

1.09

 semua

1.05

 this

1.05

Activations Density 0.165%