INDEX

Explanations

"to" followed by "how"

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

 incógn

1.25

untas

1.20

 ingrédients

1.12

 inglês

1.11

 परेशानी

1.10

bumptech

1.10

 japonés

1.09

áng

1.04

 peligro

1.04

 hongos

1.04

POSITIVE LOGITS

 如果

1.54

不

1.45

 Для

1.43

你

1.40

 对于

1.40

这

1.40

 为了

1.33

ا

1.32

학

1.31

 除了

1.30

Activations Density 0.000%