INDEX

Explanations

captives, prisoners, hostages, captured

The neuron responds to terms describing military or wartime actions (e.g. troop movements, attacks, battles).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 глоба

0.62

 abroad

0.59

Abroad

0.58

 стаби

0.57

টেও

0.56

 Eulerian

0.56

 теор

0.56

 inoxidable

0.55

膈

0.55

 гео

0.55

POSITIVE LOGITS

 garrison

1.38

 hostages

1.35

 captives

1.30

 prisoners

1.27

俘

1.27

prisoners

1.22

 artillery

1.18

 captured

1.16

 evac

1.16

 Captured

1.16

Activations Density 0.403%