INDEX

Explanations

the word "FILED" in all caps

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

 flotante

-0.77

шру

-0.77

 فريبيس

-0.74

-0.73

 HasFactory

-0.72

 ſta

-0.68

 mères

-0.66

 davran

-0.66

 ſeveral

-0.65

 Gegenteil

-0.65

POSITIVE LOGITS

FILED

2.67

 FILED

1.60

Filed

0.87

filed

0.71

 Filed

0.66

 filed

0.60

 Kirby

0.59

 filer

0.57

fil

0.56

 мар

0.55

Activations Density 0.000%