INDEX

Explanations

regulated, licensed, insured

The neuron fires on formal regulatory or licensing disclosure phrases (e.g. “authorised and regulated by…” or “licensed & insured”).

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

 until

-1.74

 when

-1.48

 затем

-1.46

Viitteet

-1.45

 then

-1.44

 ciencias

-1.42

 mitään

-1.41

unteering

-1.38

حديد

-1.37

 changes

-1.35

POSITIVE LOGITS

↵

1.32

 både

1.30

by

1.26

 berge

1.22

 kategor

1.21

 حتی

1.20

ellingham

1.20

1.18

and

1.16

 vokal

1.13

Activations Density 0.068%