INDEX

Explanations

keystone, Kerze

The neuron selectively activates on capitalized personal or brand names beginning with “Ke,” effectively spotting proper names that start with that letter sequence.

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

ኋላ

0.66

tre

0.59

સ્

0.59

Independent

0.57

 Independent

0.56

 незале

0.56

اض

0.55

 indip

0.55

 Parisi

0.54

ято

0.52

POSITIVE LOGITS

leys

0.67

వలం

0.66

 Ке

0.65

ystone

0.64

hamilan

0.61

keh

0.61

Ke

0.60

KE

0.60

जरीवाल

0.58

Ke

0.57

Activations Density 0.167%