INDEX

Explanations

regions and places

The neuron fires on words denoting named places or subdivisions—geographic regions or institutional units (e.g. “Death Row,” “Capitol Hill,” “campuses,” “wards,” etc.).

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

”,

-1.76

 Based

-1.70

></

-1.69

канал

-1.50

 вересня

-1.49

Two

-1.41

 травня

-1.40

 because

-1.40

 років

-1.38

We

-1.37

POSITIVE LOGITS

2.03

 peuvent

1.57

...'

1.54

 лучших

1.51

.'

1.45

 takiego

1.43

 sustancias

1.43

 waarop

1.42

 brauche

1.42

瓛

1.39

Activations Density 0.328%