INDEX

Explanations

research publications

New Auto-Interp

Configuration

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

verse

-0.72

soluble

-0.68

 exact

-0.66

exact

-0.59

 Portions

-0.58

 excepción

-0.57

 portions

-0.56

 along

-0.56

 fidèles

-0.56

EXACT

-0.53

POSITIVE LOGITS

 uranium

0.68

 pecuniary

0.68

 deafness

0.68

 blindness

0.66

 Uranium

0.65

 EconPapers

0.65

 gypsum

0.65

CVT

0.64

 polenta

0.63

 rheumatism

0.63

Activations Density 0.178%