INDEX

Explanations

names and legal entities

The neuron fires on formal document headings and role-or title tokens—e.g. official labels like “Trustee,” “Officer,” “ASSAULT:,” court names, judge or appellant designations—commonly found in legal opinions and police reports.

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

 لهذه

-1.67

哞

-1.66

 gethan

-1.57

 τὰ

-1.55

 zood

-1.55

だけではなく

-1.53

 yoox

-1.53

 klap

-1.53

 收納

-1.52

 oogen

-1.51

POSITIVE LOGITS

in

1.78

 polega

1.66

мл

1.63

 memanfaatkan

1.62

icolon

1.60

 По

1.59

Revenir

1.58

 There

1.58

大概是

1.57

 Because

1.52

Activations Density 0.196%