INDEX

Explanations

upper or above positions

upper followed by a descriptor

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

يع

0.58

}*/

0.57

ग्वि

0.55

 অন্ধ

0.50

 aumenta

0.48

ারা

0.48

भारतीय

0.47

 отста

0.47

vam

0.47

 individuo

0.47

POSITIVE LOGITS

 upper

1.42

 Upper

1.28

upper

1.26

 Верх

1.24

 верх

1.17

 UPPER

1.15

Upper

1.14

 above

1.13

 верхней

1.10

above

1.06

Activations Density 0.099%