INDEX

Explanations

This neuron specializes in spotting forward‐slash separators (as in paths or URLs) in the text.

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 सूत्रकृ

0.56

 अल्को

0.54

ेंद्रीय

0.50

प्टेंबर

0.49

र्यंत

0.48

 gorges

0.48

âtres

0.48

 sculptural

0.46

 ब्रिट

0.46

 فوټبال

0.46

POSITIVE LOGITS

String

0.53

['

0.50

});

0.50

})

0.49

<unused62>

0.47

OutputStream

0.47

<unused61>

0.46

="

0.46

("

0.46

Activations Density 0.196%