INDEX

Explanations

positive numbers and functions

New Auto-Interp

Configuration

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

checkable

-0.80

 inserire

-0.75

 Huey

-0.70

 uttered

-0.69

 esattamente

-0.68

ectar

-0.67

 мати

-0.67

CDS

-0.67

had

-0.66

NOK

-0.66

POSITIVE LOGITS

 nonnegative

3.44

 positive

3.36

positive

2.75

Positive

2.58

 Positive

2.50

 положи

2.17

 positiva

2.17

 positif

2.16

POSITIVE

2.14

 positivo

2.03

Activations Density 0.063%