INDEX

Explanations

class/module names followed by descriptor

This neuron detects code identifiers—especially class, service, and component names (often in CamelCase or all‐caps acronyms).

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

 ekolog

-1.42

 kolik

-1.29

 triko

-1.27

푉

-1.23

 moviliz

-1.23

 akci

-1.21

Vlast

-1.20

יפה

-1.16

 plafon

-1.16

 jeste

-1.15

POSITIVE LOGITS

 only

1.46

 will

1.43

as

1.30

in

1.28

 with

1.27

on

1.23

at

1.21

1.20

1.15

do

1.14

Activations Density 0.026%