INDEX

Explanations

official titles and roles

This neuron detects formal, document‐header terms—especially capitalized legal names, act titles, and other proper nouns found in legal filings and headings.

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

 більш

-1.49

 articol

-1.27

 forbidding

-1.26

 azzurro

-1.24

 入力

-1.22

 attesa

-1.20

 gjø

-1.20

 raí

-1.20

ELTS

-1.20

copter

-1.20

POSITIVE LOGITS

 процесс

1.28

 cerrar

1.22

</th>

1.21

 Rechten

1.20

1.15

 времени

1.13

 your

1.12

实现

1.12

ⁱ

1.11

</h4>

1.10

Activations Density 0.476%