INDEX

Explanations

leader titles followed by of

This neuron detects mentions of a meeting’s chair or presiding officer (e.g. “chair,” “head,” “president,” and their foreign-language equivalents).

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

<i>

-2.14

-2.13

-1.89

-1.83

↵↵

-1.79

."

-1.72

-1.69

,"

-1.67

-1.62

POSITIVE LOGITS

 deciso

1.88

琑

1.78

 potuto

1.74

䊊

1.73

 obiet

1.72

 lasciato

1.71

 하고

1.70

 inauguración

1.70

of

1.69

 händ

1.69

Activations Density 0.020%