INDEX

Explanations

for details or information

The neuron detects words that signal a reference or pointer to additional information (e.g. “specific,” “document,” “for,” “more,” “details,” “instructions”).

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 Unternehmen

0.67

 Enterprise

0.67

 Documents

0.59

 Whoever

0.59

 Coker

0.57

企業

0.56

 Memoir

0.56

 Thief

0.56

 whoever

0.55

 tesam

0.55

POSITIVE LOGITS

 language

0.80

 detalles

0.79

language

0.71

 limitaciones

0.71

 rattles

0.70

 details

0.70

 détails

0.68

 governing

0.67

etails

0.67

 limitations

0.66

Activations Density 0.012%