INDEX

Explanations

references to spoilage or damaging phenomena

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

 transférez

-0.59

tvguidetime

-0.59

Kontrola

-0.58

 viewDidLoad

-0.58

$/,

-0.55

 Knot

-0.54

>`;

-0.54

 disponibilités

-0.52

UTERS

-0.52

 nahilalakip

-0.52

POSITIVE LOGITS

 spoil

2.69

 spoiler

2.58

 spoiling

2.41

 spoilers

2.36

 spoils

2.20

 spoiled

2.19

Spoiler

2.14

 Spoiler

2.05

 Spoilers

1.87

spoiler

1.86

Activations Density 0.003%