INDEX

Explanations

questions and explanations

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

to

-1.90

 takich

-1.62

 related

-1.55

 Zeugen

-1.51

 cercana

-1.49

 takie

-1.48

 miest

-1.48

 такие

-1.48

なかな

-1.45

</h1>

-1.44

POSITIVE LOGITS

 fhort

1.52

"":

1.45

 neceffary

1.45

 emblema

1.43

 brillant

1.41

hua

1.41

 fufficient

1.41

 poffible

1.40

uyasha

1.37

 fhall

1.35

Activations Density 0.004%