INDEX

Explanations

telegram, war, and cryptocurrency

This neuron detects references to Russian geopolitical actions or entities (e.g. Putin, Crimea, Russian mainland, Ukraine blockade).

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

 sectarian

0.59

 Jawaharlal

0.57

сног

0.55

 Arabian

0.53

טי

0.52

 WordPerfect

0.52

ativer

0.52

gypt

0.52

 Carbohyd

0.52

 Sudan

0.52

POSITIVE LOGITS

 Zelensky

1.29

 Mariupol

1.23

 Ukrainian

1.21

 Kyiv

1.16

 Kherson

1.16

 Ukraine

1.14

 Ukrainians

1.13

Ukrainian

1.11

 Ukraj

1.09

 Volodymyr

1.09

Activations Density 0.049%