INDEX

Explanations

people killed or died

The neuron activates on numeric tokens denoting quantities—especially casualty or death counts.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

ciado

-1.32

 destinées

-1.29

 nostra

-1.27

tt

-1.27

ridae

-1.21

几种

-1.20

 adaptées

-1.19

iredo

-1.18

 nuestro

-1.14

POSITIVE LOGITS

in

1.56

ेमाल

1.55

 from

1.50

衚

1.49

КИ

1.41

 prachtig

1.39

 geweldig

1.36

 mieszkań

1.34

 έχουν

1.34

США

1.33

Activations Density 0.048%