INDEX

Explanations

(Jamaica)

The main thing this neuron does is pick out multiword named entities (proper nouns such as place names, organizations, and personal names).

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

 পশ্চিমবঙ্গের

0.77

 Ghent

0.72

Belgium

0.68

 Negros

0.67

 Ogun

0.65

Jupiter

0.64

terdam

0.64

 Rotterdam

0.63

 العراق

0.63

 Bordeaux

0.63

POSITIVE LOGITS

 Bavarian

1.41

 Bavaria

1.35

 Indore

1.34

 Madhya

1.27

 Bayern

1.26

bmw

1.26

 इंदौर

1.25

BMW

1.22

长沙

1.21

 मध्यप्रदेश

1.17

Activations Density 0.502%