INDEX

Explanations

religious texts and citations

The neuron is sensitive to named‐entity tokens and citation markers—that is, proper nouns (like personal or place names) and structured references such as book-chapter-verse or case citations.

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

in

-1.53

</h3>

-1.41

 durch

-1.35

and

-1.34

↵↵

-1.32

at

-1.27

 about

-1.24

 padx

-1.18

arxiv

-1.16

=\

-1.15

POSITIVE LOGITS

 then

2.17

Then

1.63

 télécommande

1.48

 سپس

1.41

 premières

1.40

我们就

1.37

بدأ

1.30

 faptul

1.29

start

1.28

Conclusão

1.28

Activations Density 0.002%