INDEX

Explanations

partial words, either "line", "ponentially", "haus" or "amin"

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

No Comments

Negative Logits

 farsi

-0.57

 незавершена

-0.57

drop

-0.50

 drops

-0.49

 drop

-0.49

 Exec

-0.49

zyści

-0.49

 executive

-0.48

 Executive

-0.48

commons

-0.45

POSITIVE LOGITS

 EXAMIN

0.77

Exam

0.73

蚪

0.68

 Exam

0.68

 betweenstory

0.67

 exam

0.66

 EXAM

0.66

 examiner

0.65

exam

0.65

ItemBackground

0.65

Activations Density 0.308%