INDEX

Explanations

differences and contrasts

complex systems and abstract concepts

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

 oftentimes

0.55

Ĺ

0.53

 réelle

0.52

 spotless

0.52

 اُن

0.52

🙌

0.52

⁉

0.52

 respective

0.51

 ولكن

0.51

respective

0.50

POSITIVE LOGITS

 trivially

0.84

 nontrivial

0.75

 asymptotically

0.73

 postdoc

0.72

 filesystem

0.70

 nominally

0.70

 ergodic

0.67

 plaus

0.65

 unambiguously

0.64

 equivalently

0.64

Activations Density 0.078%