INDEX

Explanations

codes, errors, and specific terminology

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

 Return

1.34

bug

1.31

nem

1.23

 airtight

1.22

nø

1.21

 return

1.20

ொள்

1.19

 nóng

1.17

niz

1.17

moc

1.17

POSITIVE LOGITS

grapes

1.59

aryl

1.50

anesu

1.48

drive

1.47

اء

1.45

 দারিদ্র

1.45

presumably

1.41

Ck

1.39

probably

1.37

পালা

1.37

Activations Density 0.000%