INDEX

Explanations

lineup

The neuron fires on mentions of an event’s “lineup” (including variants like “line-up” and “headlining”), i.e. references to who’s on the bill.

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

 different

-1.62

-1.42

ू

-1.39

↵↵

-1.35

<em>

-1.30

-1.27

 Christopher

-1.25

他人

-1.24

ktioner

-1.24

וּ

-1.23

POSITIVE LOGITS

of

1.80

''');

1.63

化作

1.48

 bordados

1.45

</b>

1.40

 plans

1.39

 bolsos

1.37

身心

1.33

鑊

1.32

jorge

1.30

Activations Density 0.033%