INDEX

Explanations

not really

descriptions and lists

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

芫

0.48

をした

0.46

 お知らせ

0.46

 Manche

0.45

 דבר

0.45

 другим

0.44

ーター

0.43

 مرات

0.43

ابس

0.43

க்கர

0.43

POSITIVE LOGITS

 homeostasis

0.49

nessy

0.47

unc

0.44

dz

0.44

 coagulation

0.42

({

0.41

heim

0.41

escape

0.41

errorHandler

0.40

-,

0.39

Activations Density 0.000%