INDEX

Explanations

patent application numbers

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

pośred

-0.94

so

-0.93

 aparato

-0.91

ironically

-0.90

 gyerme

-0.90

 stratég

-0.88

corrosion

-0.87

 gorro

-0.85

もあるので

-0.85

 lourdes

-0.84

POSITIVE LOGITS

autres

0.91

 calend

0.87

事例

0.87

 escarp

0.85

 definitiva

0.84

”;

0.82

hell

0.82

 verboten

0.81

NYU

0.80

to

0.78

Activations Density 0.077%