INDEX

Explanations

pure, clean

pure or clean followed by concepts

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

 adecuada

0.70

 fopen

0.70

 desigualdad

0.63

 controla

0.62

 []);

0.61

 paysans

0.61

 sayıda

0.61

 нижней

0.60

 Puede

0.59

 ayaa

0.59

POSITIVE LOGITS

 pure

1.27

Pure

1.24

pure

1.22

 Pure

1.19

纯

1.16

PURE

1.13

純

1.06

 чисто

1.03

 purely

0.95

 शुद्ध

0.91

Activations Density 0.092%