INDEX

Explanations

protecting, improving, confirming

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 тако

1.16

обре

1.13

fois

1.07

dic

0.99

 مجبور

0.98

 leggi

0.97

 некоторым

0.96

 clerks

0.96

 thus

0.95

dh

0.94

POSITIVE LOGITS

 pointB

1.35

 POINT

1.33

ೋತಿ

1.31

POINT

1.31

 포인트

1.25

 μικ

1.25

 Point

1.22

мын

1.22

 maximizing

1.21

焦點

1.20

Activations Density 0.441%