INDEX

Explanations

specific compound phrases

no access to real-time information

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 বাড়ী

0.64

 गाँधी

0.62

 বেশী

0.61

 পৌছ

0.61

 atleast

0.61

&/

0.59

 তৈরী

0.57

 জরুরী

0.55

 இவ்வ

0.54

 গ্রহন

0.54

POSITIVE LOGITS

 굉장히

1.15

すごく

1.13

1.05

 väldigt

1.05

比如说

1.05

いろんな

1.02

 그다음에

0.95

 veldig

0.94

 really

0.92

 خیلی

0.86

Activations Density 0.087%