INDEX

Explanations

ensuring safety and correctness

ensuring accuracy and reliability

New Auto-Interp

Configuration

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

Negative Logits

 менее

0.41

 лично

0.40

también

0.40

зже

0.39

ítez

0.39

ärt

0.39

并非

0.39

твор

0.39

かもしれない

0.39

㳡

0.39

POSITIVE LOGITS

 ensure

0.70

 ensuring

0.69

 ensures

0.67

 perfekt

0.67

 ensured

0.65

 zapewn

0.64

 perfetto

0.63

 đảm

0.62

 memastikan

0.62

 excellently

0.61

Activations Density 0.035%