INDEX

Explanations

reject or may not

boilerplate generation

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

ケーション

0.46

 deviation

0.45

apses

0.44

それが

0.43

足り

0.43

 modific

0.41

 obliterated

0.41

 discard

0.39

 tweaking

0.39

 tweaked

0.39

POSITIVE LOGITS

steppe

0.45

তর

0.44

 Bomber

0.42

”

0.41

 chum

0.41

 Mena

0.41

 તેમ

0.40

Vib

0.40

 XXIII

0.40

 njemu

0.40

Activations Density 0.000%