INDEX

Explanations

former or current officeholders

The neuron detects temporal or status qualifiers attached to people or events (e.g. “outgoing,” “former,” “current,” “then,” “new,” “incumbent”).

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

ième

-1.63

횃

-1.44

 sesama

-1.41

Those

-1.40

itself

-1.39

але

-1.39

Being

-1.38

///</

-1.34

 конструкции

-1.30

 штук

-1.30

POSITIVE LOGITS

who

2.22

of

2.03

 president

1.95

 whom

1.74

1.51

prodi

1.43

 father

1.41

playstation

1.38

 there

1.38

 doesn

1.37

Activations Density 0.030%