INDEX

Explanations

confirming correctness or meaning

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 لكن

1.72

but

1.71

 albeit

1.69

 though

1.65

 אבל

1.65

though

1.63

but

1.60

だけど

1.59

although

1.55

 although

1.54

POSITIVE LOGITS

 माननीय

1.37

جميع

1.21

 dignitaries

1.21

株式会社

1.21

 congratulated

1.19

 commended

1.17

இந்

1.17

 उपस्थित

1.16

 كافة

1.16

 জনাব

1.15

Activations Density 0.161%