INDEX

Explanations

phrases with specific completions

initial assessment

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

 carcinomas

0.50

 overpower

0.49

 केंद्रों

0.46

 evidencia

0.45

 całkow

0.44

 weighed

0.44

 biodegrad

0.44

 indestruct

0.44

 overwhelm

0.43

 eliminate

0.43

POSITIVE LOGITS

ون

0.52

jom

0.47

apen

0.46

נס

0.45

מת

0.45

хими

0.45

ᵈ

0.44

ap

0.44

JIM

0.43

DATA

0.43

Activations Density 0.000%