INDEX

Explanations

UK, Britain, England

The neuron activates on multi‐word named entities (proper nouns)—places, organizations, roads, people’s names, etc.

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

 marijuana

1.38

 NASCAR

1.37

 demeanor

1.34

 amerik

1.32

 Galveston

1.30

 Tahoe

1.29

 cognizant

1.28

 ACLU

1.27

 Tulane

1.27

 shortstop

1.27

POSITIVE LOGITS

 британ

2.67

British

2.63

 Великобритании

2.59

英国

2.58

Britain

2.56

英國

2.55

London

2.52

England

2.50

Leicester

2.48

イギリス

2.47

Activations Density 3.790%