INDEX

Explanations

correctness and alternative phrasing

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

 belirl

1.30

 connaissances

1.21

୯

1.16

 conocimiento

1.14

 знания

1.12

Ⅷ

1.10

ා

1.10

}}\|

1.09

 знаний

1.09

 conoscenze

1.09

POSITIVE LOGITS

 would

1.06

 seems

1.00

 simpler

0.94

 toast

0.94

 seeming

0.94

 implying

0.94

 defies

0.93

 nicer

0.93

 louder

0.92

 translates

0.92

Activations Density 0.122%