INDEX

Explanations

error traceback recover crash

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 whom

-1.68

 kurios

-1.52

ссия

-1.38

我们

-1.36

我們

-1.35

因為

-1.34

.《

-1.30

æld

-1.29

簠

-1.26

 EACH

-1.25

POSITIVE LOGITS

’

1.56

“

1.50

 conocida

1.43

↵

1.38

1.30

 όπου

1.30

そこから

1.30

 meander

1.29

情况下

1.25

气势

1.24

Activations Density 0.043%