INDEX

Explanations

equations and technical terms

The neuron consistently lights up on the main content word at the end of a clause or sentence—typically the sentence-final noun or verb that carries its semantic “head.”

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

0.72

Repository

0.69

 Operational

0.69

 vices

0.68

 действие

0.68

ium

0.67

 operational

0.67

 senses

0.67

ד

0.67

 continuously

0.66

POSITIVE LOGITS

 स्प्रे

0.88

 धमाकेदार

0.85

}_{\

0.84

ტის

0.79

 सेटिंग

0.76

参照

0.75

 तरल

0.75

andaş

0.75

 echter

0.74

 शुगर

0.73

Activations Density 0.000%