INDEX

Explanations

research papers

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ed

-0.68

listdir

-0.56

 touched

-0.53

 Complementary

-0.52

touched

-0.51

ived

-0.49

 convective

-0.47

ized

-0.47

ibus

-0.47

POSITIVE LOGITS

 OFDb

0.79

 ويكيپيديا

0.65

úgó

0.63

 باخ

0.58

נחנו

0.56

 متعلقه

0.56

énario

0.55

 nahilalakip

0.55

 XCTest

0.54

aneous

0.54

Activations Density 0.059%