INDEX

Explanations

demonstrates usage and provides feedback

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 الناس

1.40

 اسم

1.32

 люди

1.31

 مثل

1.31

ي

1.30

 человеку

1.28

 ولا

1.27

 людей

1.26

sail

1.26

 спине

1.25

POSITIVE LOGITS

can

1.22

 must

1.03

 should

0.97

 می‌توان

0.96

 complementar

0.95

 blends

0.95

 kits

0.95

 erreichen

0.93

attachment

0.92

 neutralizing

0.92

Activations Density 0.263%