INDEX

Explanations

the fact that

judgments and numerical values

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

ామ

0.50

ศึก

0.45

hor

0.44

 fabri

0.43

nom

0.43

 coron

0.43

 Cidade

0.42

 beast

0.41

IND

0.40

짐

0.40

POSITIVE LOGITS

ل

0.59

ল

0.57

ク

0.55

л

0.52

 invariably

0.49

nicheskij

0.49

 අවශ්‍ය

0.49

ن

0.48

chés

0.48

 necesariamente

0.48

Activations Density 0.000%