INDEX

Explanations

`public class`

The neuron flags bibliographic‐style metadata tokens—especially numbers (page ranges, years), place names (e.g. publishers’ locations), and other reference codes.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 joten

0.54

 場合

0.50

 দায়

0.49

 plej

0.48

agm

0.46

 ため

0.46

ष्टिक

0.46

-/

0.46

<0xD0>

0.46

 espécie

0.45

POSITIVE LOGITS

3.53

：

2.33

:\

2.33

:"

2.25

:}

2.25

:-

2.23

:'

2.22

:',

2.14

:,

2.09

 :");

1.96

Activations Density 0.061%