INDEX

Explanations

Maria followed by names

The neuron consistently activates on the proper name “Maria” (and its variants) whenever it appears.

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

 follic

-0.94

 domani

-0.90

 gesagt

-0.89

 χωρίς

-0.88

 الخارجية

-0.87

 gelegt

-0.86

Actually

-0.84



-0.81

 ricordi

-0.81

 موب

-0.79

POSITIVE LOGITS

 Elena

1.02

elfde

0.96

 Montessori

0.93

vantaged

0.89

 Luisa

0.88

 Theresa

0.87

 Magdalena

0.85

Elena

0.85

edere

0.84

Лу

0.84

Activations Density 0.010%