INDEX

Explanations

the letter 'j' in foreign (likely Slavic) language text

j

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

Negative Logits

-1.59

-1.15

-0.84

ju

-0.76

-0.67

jo

-0.61

 place

-0.59

</b>

-0.57

jj

-0.56

POSITIVE LOGITS

 '\\;'

0.89

findpost

0.76

ValueStyle

0.73

FieldNumber

0.71

новниш

0.71

bbene

0.70

StructEnd

0.68

RetentionPolicy

0.68

berdayakan

0.67

anskje

0.67

Activations Density 0.536%