INDEX

Explanations

first-person pronouns referencing questions or statements in different domains (beekeeping, medicine, DNS records, loans, and speech recognition).

comma

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

Negative Logits

-0.46

asiest

-0.43

arxiv

-0.39

everywhere

-0.38

Upstairs

-0.37

 mañana

-0.36

the

-0.36

TRIES

-0.36

 Nowak

-0.35

asi

-0.35

POSITIVE LOGITS

IntoConstraints

0.82

 الحره

0.81

 EconPapers

0.79

+#+

0.77

 wherefore

0.74

0.70

 ivelany

0.70

msgTypes

0.70

 nahilalakip

0.66

tagHelperRunner

0.65

Activations Density 2.638%