INDEX

Explanations

indication

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 individual

-1.24

 indication

-1.16

individual

-1.13

 individuales

-1.05

 Giappone

-0.99

 japonais

-0.96

 giapp

-0.94

 japoneses

-0.94

Individual

-0.94

 individuale

-0.92

POSITIVE LOGITS

ization

0.67

Portale

0.61

.*")]

0.59

’

0.58

ised

0.56

 that

0.55

ized

0.53

ban

0.52

izing

0.51

Activations Density 1.037%