INDEX

Explanations

2008 presidential election

The neuron detects numeric date references, especially years (e.g. “2008,” “2007,” “2010”).

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

ቍ

-1.52

𞤢

-1.34

赑

-1.34

 pelear

-1.32

 lenovo

-1.32

 развитии

-1.31

＾＾；

-1.30

isição

-1.30

 amaze

-1.29

یات

-1.28

POSITIVE LOGITS

of

1.37

—

1.34

ourcen

1.30

would

1.25

 writes

1.25

 rapidement

1.23

 metre

1.23

 rumour

1.23

 celebrations

1.21

 pajamas

1.21

Activations Density 0.093%