INDEX

Explanations

numbers and slashes in code

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

сне

-1.02

ujen

-0.98

சி

-0.96

 Interessen

-0.96

 APPELLANT

-0.92

FormBorderStyle

-0.90

atoon

-0.88

 técnico

-0.88

อ่าน

-0.87

 ليست

-0.87

POSITIVE LOGITS

 reputable

0.98

 הָ

0.90

 directing

0.90

 Propri

0.84

 integrating

0.83

 indicating

0.82

 unequivocally

0.82

 granic

0.82

いら

0.81

ᅧ

0.81

Activations Density 0.002%