INDEX

Explanations

desire or willingness to act

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

лло

1.34

เคย

1.32

юра

1.32

 Gazetteer

1.25

ηγ

1.20

лган

1.17

涕

1.13

дено

1.11

dessus

1.10

โมง

1.08

POSITIVE LOGITS

 graag

1.59

 möglichst

1.58

 tahu

1.57

 sabe

1.50

 desperately

1.47

<unused260>

1.47

 replace

1.47

 badly

1.45

<unused398>

1.45

ரிய

1.45

Activations Density 0.974%