INDEX

Explanations

pronoun followed by verb or article

account compromised

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

outing

0.47

esh

0.46

crocodile

0.46

 Inflammation

0.44

 Faroe

0.43

 plexus

0.42

 Leistungs

0.42

embal

0.42

 Peaceful

0.42

ोग्राफी

0.41

POSITIVE LOGITS

，

0.64

 adopts

0.56

،

0.56

nú

0.52

）,

0.50

），

0.50

 primitives

0.48

 revenues

0.47

 cach

0.46

呚

0.46

Activations Density 0.001%