INDEX

Explanations

references to thorough and all-encompassing analyses or evaluations

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 disambiguazione

-0.59

 licence

-0.57

 Evol

-0.57

 harves

-0.55

TagMode

-0.55

 Inversion

-0.54

wohner

-0.53

semitism

-0.53

 Licence

-0.52

atoday

-0.52

POSITIVE LOGITS

Comprehensive

2.72

 comprehensive

2.69

 Comprehensive

2.66

comprehensive

2.55

 detailed

1.88

 Detailed

1.79

prehensive

1.75

 thorough

1.73

Detailed

1.69

 comprehensively

1.69

Activations Density 0.037%