INDEX

Explanations

DOIs and ISBNs

The neuron fires on numeric identifiers and values (digit sequences, often with decimals) appearing after labels like “ISSN:”, “Pages:”, “DOI:”, “ISBN:”, dates, etc.

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

If

-2.09

-1.77

When

-1.68

了一

-1.65

ًا

-1.63

 insisted

-1.59

Its

-1.57

If

-1.51

are

-1.50

 Caused

-1.50

POSITIVE LOGITS

珦

1.82

潆

1.76

桄

1.65

 encontraban

1.43

 gier

1.41

 vivía

1.35

 示例

1.33

 Великобрита

1.33

蒾

1.30

 jaka

1.29

Activations Density 0.038%