INDEX

Explanations

rules, policy, plan, protocol, manual

The neuron activates on formal document or section titles—single words like “Plan,” “Resolution,” “Rule,” “Policy,” “Protocol,” “Statement,” or “Guide” that denote official headings in texts.

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

 solche

-1.13

 conocidas

-1.05

 مثل

-1.05

 conocida

-1.02

-1.00

Sunday

-0.98

↵

-0.98

<b>

-0.97

-0.96

POSITIVE LOGITS

 biela

1.45

 which

1.33

and

1.27

 rése

1.26

 vertes

1.24

of

1.22

erobic

1.18

 parfüm

1.17

 bús

1.16

 kering

1.16

Activations Density 0.033%