INDEX

Explanations

North Korea

The neuron strongly activates on mentions of “North Korea” (and closely related tokens like “Korea” or “Korean”), effectively identifying references to that country.

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

-1.82

-1.52

-1.51

ll

-1.50

da

-1.50

\},

-1.45

Also

-1.44

-1.43

"";

-1.41

POSITIVE LOGITS

to

2.13

 seleccione

1.41

ка

1.40

</b>

1.38

is

1.37

ſten

1.34

standes

1.33

mässig

1.32

ÍAS

1.31

 cuarta

1.30

Activations Density 0.007%