INDEX

Explanations

again

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

 mehrere

-0.94

 erste

-0.92

 manchmal

-0.89

一些

-0.88

Often

-0.86

 иногда

-0.85

 verschiedene

-0.83

 sogenannten

-0.82

 ersten

-0.82

ATURES

-0.80

POSITIVE LOGITS

 again

4.03

again

2.94

Again

2.69

 Again

2.67

 similar

2.48

 nuevamente

2.45

 novamente

2.39

 opět

2.39

 AGAIN

2.38

 yine

2.28

Activations Density 0.033%