INDEX

Explanations

identifying role or objective

importance or choice

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

 farthest

0.40

 painstakingly

0.40

 brilliantly

0.39

 furthest

0.39

と思っていた

0.38

 generously

0.38

contrast

0.38

 breezy

0.38

 diminutive

0.37

 eponymous

0.37

POSITIVE LOGITS

 onus

0.65

 role

0.61

 responsibility

0.61

 essence

0.58

aim

0.57

 ROLE

0.56

 term

0.55

 objectives

0.54

 adecuada

0.54

 aims

0.54

Activations Density 0.065%