INDEX

Explanations

conformity and deviation

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

ausea

-0.89

δω

-0.89

alny

-0.85

 intestino

-0.83

ἐ

-0.83

 kvalit

-0.81

Качество

-0.80

銖

-0.80

 pulumi

-0.80

tså

-0.80

POSITIVE LOGITS

 conform

2.19

 conforming

2.16

 conformity

2.05

 conforms

2.05

conform

1.93

 norms

1.92

 conventional

1.91

 norm

1.88

conventional

1.76

 Conform

1.72

Activations Density 0.047%