INDEX

Explanations

suffix or non-English word

number of parameters

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Negative Logits

inputId

0.49

ън

0.44

eventId

0.44

τε

0.43

 Task

0.40

街

0.40

问题的

0.40

问题

0.39

 задачу

0.39

нат

0.39

POSITIVE LOGITS

istischen

0.46

ynn

0.45

aceae

0.44

swire

0.44

baix

0.44

 পুনরায়

0.42

icom

0.42

pecial

0.42

 दुर्लभ

0.42

ន្ថ

0.42

Activations Density 0.001%